A comparison of maximum covariance and K-means cluster analysis in classifying cases into known taxon groups

Assessment of college students' mental health status based on temporal perception and hybrid clustering algorithm under the impact of public health events

The dynamic landscape of public health occurrences presents a formidable challenge to the emotional well-being of college students, necessitating a precise appraisal of their mental health (MH) status. A pivotal metric in this realm is the Mental Health Assessment Index, a prevalent gauge utilized to ascertain an individual's psychological well-being. However, prevailing indices predominantly stem from a physical vantage point, neglecting the intricate psychological dimensions. In pursuit of a judicious evaluation of college students' mental health within the crucible of public health vicissitudes, we have pioneered an innovative metric, underscored by temporal perception, in concert with a hybrid clustering algorithm. This augmentation stands poised to enrich the extant psychological assessment index framework. Our approach hinges on the transmutation of temporal perception into a quantifiable measure, harmoniously interwoven with established evaluative metrics, thereby forging a novel composite evaluation metric. This composite metric serves as the fulcrum upon which we have conceived a pioneering clustering algorithm, seamlessly fusing the fireworks algorithm with K-means clustering. The strategic integration of the fireworks algorithm addresses a noteworthy vulnerability inherent to K-means-its susceptibility to converging onto local optima. Empirical validation of our paradigm attests to its efficacy. The proposed hybrid clustering algorithm aptly captures the dynamic nuances characterizing college students' mental health trajectories. Across diverse assessment stages, our model consistently attains an accuracy threshold surpassing 90%, thus outshining existing evaluation techniques in both precision and simplicity. In summation, this innovative amalgamation presents a formidable stride toward an augmented understanding of college students' mental well-being during times of fluctuating public health dynamics.

Effect of Tensile Frequency on the Osteogenic Differentiation of Periodontal Ligament Stem Cells

Purpose

The role of periodontal ligament stem cells (PDLSCs) in mediating osteogenesis involved in orthodontic tooth movement (OTM) is well established. However, various relevant in vitro studies vary in the frequency of tension. The effect of tensile frequency on the mechanotransduction of PDLSCs is not clear. The current study aimed to determine the effect of different tensile frequencies on the osteogenic differentiation of PDLSCs and to identify important mechano-sensitivity genes.

Methods

Human PDLSCs were isolated, identified, and subjected to cyclic equibiaxial tensile strain of 12% at different frequencies of 0.1 Hz, 0.5 Hz, 0.7 Hz, or static cultures. Osteogenic differentiation of PDLSCs was assessed by using Western blotting. High-throughput sequencing was used to identify differential mRNA expression. Short time-series expression miner (STEM) was utilized to describe the frequency patterns of the mRNAs. The functions and enriched pathways were identified, and the hub genes were identified and validated.

Results

We found that the osteoblastic differentiation capacity of PDLSCs increased with tensile frequency in the range of 0.1–0.7 Hz. Eight frequency-tendency gene expression profiles were identified to be statistically significant. Tensile frequency-specific expressed genes, such as SALL1 and EYA1, which decreased with the increase in tensile frequency, were found.

Conclusion

The osteoblastic differentiation of PDLSCs under mechanical tensile force is frequency dependent. EYA1 and SALL1 were identified as potential important tensile frequency-sensitive genes, which may contribute to the cyclic tension-induced osteogenic differentiation of PDLSCs in a frequency-dependent manner.

Food addiction symptoms are related to neuroaffective responses to preferred binge food and erotic cues

It has been proposed that some individuals succumb to maladaptive eating behaviors because, like those with addiction, they attribute high incentive salience to food-associated cues. Here, we tested whether women that attribute high incentive salience to food-associated cues report high food addiction symptomatology. In 76 college women, we assessed self-reported food addiction symptoms using the Yale Food Addiction Scale and we recorded event-related potentials (ERPs, a direct measure of brain activity) to preferred food, erotic, unpleasant, and neutral images. We used the amplitude of the late positive potential (LPP, a component of the ERPs) as an index of the incentive salience attributed to the images. Using a multivariate classification algorithm (k-means cluster analysis), we identified two neuroaffective reactivity profiles that have been previously associated with individual differences in the tendency to attribute incentive salience to cues and with differences in vulnerability to addictive behaviors. Results showed that women with elevated LPP responses to preferred food cues relative to erotic images report higher food addiction symptoms than women with low LPP responses to preferred food cues relative to other motivationally relevant stimuli. These results support the hypothesis that individual differences in the tendency to attribute incentive salience to food cues play an important role in modulating food addiction symptomatology.

A Monte Carlo evaluation of growth mixture modeling

Growth mixture modeling (GMM) and its variants, which group individuals based on similar longitudinal growth trajectories, are quite popular in developmental and clinical science. However, research addressing the validity of GMM-identified latent subgroupings is limited. This Monte Carlo simulation tests the efficiency of GMM in identifying known subgroups (k = 1-4) across various combinations of distributional characteristics, including skew, kurtosis, sample size, intercept effect size, patterns of growth (none, linear, quadratic, exponential), and proportions of observations within each group. In total, 1,955 combinations of distributional parameters were examined, each with 1,000 replications (1,955,000 simulations). Using standard fit indices, GMM often identified the wrong number of groups. When one group was simulated with varying skew and kurtosis, GMM often identified multiple groups. When two groups were simulated, GMM performed well only when one group had steep growth (whether linear, quadratic, or exponential). When three to four groups were simulated, GMM was effective primarily when intercept effect sizes and sample sizes were large, an uncommon state of affairs in real-world applications. When conditions were less ideal, GMM often underestimated the correct number of groups when the true number was between two and four. Results suggest caution in interpreting GMM results, which sometimes get reified in the literature.

Sensor-based categorization of upper limb performance in daily life of persons with and without neurological upper limb deficits

Background: The use of wearable sensor technology (e. g., accelerometers) for tracking human physical activity have allowed for measurement of actual activity performance of the upper limb (UL) in daily life. Data extracted from accelerometers can be used to quantify multiple variables measuring different aspects of UL performance in one or both limbs. A limitation is that several variables are needed to understand the complexity of UL performance in daily life.

Purpose: To identify categories of UL performance in daily life in adults with and without neurological UL deficits.

Methods: This study analyzed data extracted from bimanual, wrist-worn triaxial accelerometers from adults from three previous cohorts (N = 211), two samples of persons with stroke and one sample from neurologically intact adult controls. Data used in these analyses were UL performance variables calculated from accelerometer data, associated clinical measures, and participant characteristics. A total of twelve cluster solutions (3-, 4-, or 5-clusters based with 12, 9, 7, or 5 input variables) were calculated to systematically evaluate the most parsimonious solution. Quality metrics and principal component analysis of each solution were calculated to arrive at a locally-optimal solution with respect to number of input variables and number of clusters.

Results: Across different numbers of input variables, two principal components consistently explained the most variance. Across the models with differing numbers of UL input performance variables, a 5-cluster solution explained the most overall total variance (79%) and had the best model-fit.

Conclusion: The present study identified 5 categories of UL performance formed from 5 UL performance variables in cohorts with and without neurological UL deficits. Further validation of both the number of UL performance variables and categories will be required on a larger, more heterogeneous sample. Following validation, these categories may be used as outcomes in UL stroke research and implemented into rehabilitation clinical practice.

The classification of eating disorders in China: A categorical model or a dimensional model

OBJECTIVE

According to the ICD-10 and DSM-5, eating disorders (EDs) are classified using a categorical model that assumes the subtypes are qualitatively different from one another. However, it is still intensely debated that a dimensional model is more suitable. The aim of this study is to examine whether EDs have a categorical or dimensional latent structure using a sample of Chinese ED patients.

METHOD

The sample included 322 patients, diagnosed with an ED from 2010 to 2017 in the Shanghai Mental Health Center, and comparison participants (N = 850), recruited from undergraduate students in one university in Shanghai. Participants were evaluated with the Eating Disorder Inventory-2 (EDI-2) questionnaire and another questionnaire developed by the researchers. Three taxometric procedures (MAXimum EIGenvalue [MAXEIG], latent-mode factor analysis [L-Mode], and Mean Above Minus Below A Cut [MAMBAC]) were applied, respectively, to analyze the patients' clinical symptoms data.

RESULTS

Patients were divided into three groups according to their clinical diagnosis. The plots of the three taxometric analysis procedures supported the categorical construct in anorexia nervosa, binge-eating/purging group, and bulimia nervosa group. The Comparison Curve Fit Indices of the MAXEIG, L-Mode, and MAMBAC procedures were 0.694, 0.709, 0.704 in the AN-BP group and 0.727, 0.67, 0.62 in the BN group, respectively, which also support the categorical construct.

DISCUSSION

The results support two distinct classes of ED subtypes among Chinese sample. Further work on applying hybrid model in analysis has been discussed.

Do recognizable lifetime eating disorder phenotypes naturally occur in a culturally asian population? A combined latent profile and taxometric approach

BACKGROUND

We examined whether empirically derived eating disorder (ED) categories in Hong Kong Chinese patients (N = 454) would be consistent with recognizable lifetime ED phenotypes derived from latent structure models of European and American samples.

METHOD

We performed latent profile analysis (LPA) using indicator variables from data collected during routine assessment, and then applied taxometric analysis to determine whether latent classes were qualitatively versus quantitatively distinct.

RESULTS

Latent profile analysis identified four classes: (i) binge/purge (47%); (ii) non-fat-phobic low-weight (34%); (iii) fat-phobic low-weight (12%); and (iv) overweight disordered eating (6%). Taxometric analysis identified qualitative (categorical) distinctions between the binge/purge and non-fat-phobic low-weight classes, and also between the fat-phobic and non-fat-phobic low-weight classes. Distinctions between the fat-phobic low-weight and binge/purge classes were indeterminate.

CONCLUSION

Empirically derived categories in Hong Kong showed recognizable correspondence with recognizable lifetime ED phenotypes. Although taxometric findings support two distinct classes of low weight EDs, LPA findings also support heterogeneity among non-fat-phobic individuals.

Revealing gene clusters associated with the development of cholangiocarcinoma, based on a time series analysis

Cholangiocarcinoma (CC) is a rapidly lethal malignancy and currently is considered to be incurable. Biomarkers related to the development of CC remain unclear. The present study aimed to identify differentially expressed genes (DEGs) between normal tissue and intrahepatic CC, as well as specific gene expression patterns that changed together with the development of CC. By using a two‑way analysis of variance test, the biomarkers that could distinguish between normal tissue and intrahepatic CC dissected from different days were identified. A k‑means cluster method was used to identify gene clusters associated with the development of CC according to their changing expression pattern. Functional enrichment analysis was used to infer the function of each of the gene sets. A time series analysis was constructed to reveal gene signatures that were associated with the development of CC based on gene expression profile changes. Genes related to CC were shown to be involved in 'mitochondrion' and 'focal adhesion'. Three interesting gene groups were identified by the k‑means cluster method. Gene clusters with a unique expression pattern are related with the development of CC. The data of this study will facilitate novel discoveries regarding the genetic study of CC by further work.

DSM-IV antisocial personality disorder and conduct disorder: evidence for taxonic structures among individuals with and without substance use disorders in the general population

OBJECTIVE

The categorical-dimensional status of DSM-IV (Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition) conduct disorder (CD) and antisocial personality disorder (ASPD) is a source of controversy. This study examined whether the underlying structure of DSM-IV CD and ASPD was dimensional or categorical (taxonic) among individuals with and without substance use disorders.

METHOD

Using a national large representative survey of U.S. adults (n = 43,093), taxometric analyses of DSM-IV CD and ASPD diagnostic criteria were conducted on the total sample and among those with and without substance use disorders.

RESULTS

Results of three taxometric procedures were consistent in showing that the structures underlying DSM-IV CD and ASPD were clearly taxonic in the total sample and among individuals with and without substance use disorders. Comparison curve fit indices exceeded 0.57 for each model.

CONCLUSIONS

Taxonic findings of the present study were in contrast to the dimensional results of prior taxometric research among incarcerated samples with substantial comorbidity of antisocial syndromes and substance use disorders. Results supported the categorical representation and diagnostic thresholds of ASPD and CD as defined in DSM-IV and DSM-5. That the structure of ASPD and CD may be taxonic suggests that further research on these disorders use group comparative designs in which samples with and without these disorders are compared in terms of sociodemographic and clinical correlates, comorbidity, and treatment utilization. The taxonic structure of ASPD and CD may contribute to future research on causal processes through which these antisocial syndromes develop.

DSM-IV schizotypal personality disorder: a taxometric analysis among individuals with and without substance use disorders in the general population

This study examined the underlying structure of DSM-IV schizotypal personality disorder (SPD) among individuals with and without a substance use disorder. Using a nationally representative sample of U.S. adults, taxometric analyses were conducted on SPD in the total sample and among individuals with and without a substance use disorder. The structure of SPD in the total sample and among individuals without substance use disorders was dimensional (comparison curve fit indices (CCFI): 0.440, 0.365) whereas a taxonic structure was demonstrated among individuals with a substance use disorder (CCFI: 0.679). Taxonicity underlying schizotypy and SPD in prior taxometric research may have been the result of sampling high risk subsamples of the population. Taxometric research on SPD and other personality psychopathology among high risk subgroups of the population can help elucidate the complex etiology of SPD and the role played by comorbid substance use disorders in the expressivity of these disorders.

Group membership prediction when known groups consist of unknown subgroups: a Monte Carlo comparison of methods

Classification using standard statistical methods such as linear discriminant analysis (LDA) or logistic regression (LR) presume knowledge of group membership prior to the development of an algorithm for prediction. However, in many real world applications members of the same nominal group, might in fact come from different subpopulations on the underlying construct. For example, individuals diagnosed with depression will not all have the same levels of this disorder, though for the purposes of LDA or LR they will be treated in the same manner. The goal of this simulation study was to examine the performance of several methods for group classification in the case where within group membership was not homogeneous. For example, suppose there are 3 known groups but within each group two unknown classes. Several approaches were compared, including LDA, LR, classification and regression trees (CART), generalized additive models (GAM), and mixture discriminant analysis (MIXDA). Results of the study indicated that CART and mixture discriminant analysis were the most effective tools for situations in which known groups were not homogeneous, whereas LDA, LR, and GAM had the highest rates of misclassification. Implications of these results for theory and practice are discussed.

Determining the Number of Latent Classes in Single- and Multi-Phase Growth Mixture Models

Stage-sequential (or multiphase) growth mixture models are useful for delineating potentially different growth processes across multiple phases over time and for determining whether latent subgroups exist within a population. These models are increasingly important as social behavioral scientists are interested in better understanding change processes across distinctively different phases, such as before and after an intervention. One of the less understood issues related to the use of growth mixture models is how to decide on the optimal number of latent classes. The performance of several traditionally used information criteria for determining the number of classes is examined through a Monte Carlo simulation study in single- and multi-phase growth mixture models. For thorough examination, the simulation was carried out in two perspectives: the models and the factors. The simulation in terms of the models was carried out to see the overall performance of the information criteria within and across the models, while the simulation in terms of the factors was carried out to see the effect of each simulation factor on the performance of the information criteria holding the other factors constant. The findings not only support that sample size adjusted BIC (ADBIC) would be a good choice under more realistic conditions, such as low class separation, smaller sample size, and/or missing data, but also increase understanding of the performance of information criteria in single- and multi-phase growth mixture models.

Taxometric analysis of DSM-IV and DSM-5 alcohol use disorders

BACKGROUND

With preparations currently being made for the Diagnostic and Statistical Manual of Mental Disorders-5th Edition (DSM-5), one prominent issue to resolve is whether alcohol use disorders are better represented as discrete categorical entities or as a dimensional construct. The purpose of this study was to investigate the latent structure of DSM-4th edition (DSM-IV) and proposed DSM-5 alcohol use disorders.

METHODS

The study used the Wave 2 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) to conduct taxometric analyses of DSM-IV and DSM-5 alcohol use disorders defined by different thresholds to determine the taxonic or dimensional structure underlying the disorders.

RESULTS

DSM-IV and DSM-5 alcohol abuse and dependence criteria with 3+ thresholds demonstrated a dimensional structure. Corresponding thresholds with 4+ criteria were clearly taxonic, as were thresholds defined by cut-offs of 5+ and 6+ criteria.

CONCLUSIONS

DSM-IV and DSM-5 alcohol use disorders demonstrated a hybrid taxonic-dimensional structure. That is, DSM-IV and DSM-5 alcohol use disorders may be taxonically distinct compared to no disorder if defined by a threshold of 4 or more criteria. However, there may be dimensional variation remaining among non-problematic to subclinical cases. A careful and systematic program of structural research using taxometric and psychometric procedures is warranted.

KmL3D: a non-parametric algorithm for clustering joint trajectories

In cohort studies, variables are measured repeatedly and can be considered as trajectories. A classic way to work with trajectories is to cluster them in order to detect the existence of homogeneous patterns of evolution. Since cohort studies usually measure a large number of variables, it might be interesting to study the joint evolution of several variables (also called joint-variable trajectories). To date, the only way to cluster joint-trajectories is to cluster each trajectory independently, then to cross the partitions obtained. This approach is unsatisfactory because it does not take into account a possible co-evolution of variable-trajectories. KmL3D is an R package that implements a version of k-means dedicated to clustering joint-trajectories. It provides facilities for the management of missing values, offers several quality criteria and its graphic interface helps the user to select the best partition. KmL3D can work with any number of joint-variable trajectories. In the restricted case of two joint trajectories, it proposes 3D tools to visualize the partitioning and then export 3D dynamic rotating-graphs to PDF format.

Exploring the clinical validity of borderline personality disorder components

Borderline personality disorder (BPD) is recognized as a complex syndrome, resulting in a heterogeneous diagnostic category. Besides the characteristics of the disorder itself, comorbid disorders play an important role in this complexity. The aim of the study is to analyze the clinical validity of 3 components for BPD Diagnostic and Statistical Manual of Mental Disorders criteria--called affective dysregulation, behavioral dysregulation, and disturbed relatedness--investigating differences in patterns of comorbidity. For this purpose, 365 patients with suspected BPD were included in the study. To test our hypothesis, patients were classified into 5 clusters using a K-cluster analysis to study the clinical validity of the 3 components based on the 3-factor model of BPD. Differences in comorbidity, previous suicide attempts, and self-harm behaviors among the defined clusters were analyzed. Between-cluster differences were observed for Axis I and Axis II disorders as well as in the frequency of suicide attempts and in self-harm behaviors. The study of BPD based on the 3 components seems to be more useful than the study of BPD as a unitary construct to help further our understanding of this complex disorder. In the present study, the 3 BPD components have allowed us to analyze the complex comorbidity of BPD patients. This solution could be considered an interesting way to clarify BPD etiology, diagnosis, and treatment efficacy.

KmL: a package to cluster longitudinal data

Cohort studies are becoming essential tools in epidemiological research. In these studies, measurements are not restricted to single variables but can be seen as trajectories. Thus, an important question concerns the existence of homogeneous patient trajectories. KmL is an R package providing an implementation of k-means designed to work specifically on longitudinal data. It provides several different techniques for dealing with missing values in trajectories (classical ones like linear interpolation or LOCF but also new ones like copyMean). It can run k-means with distances specifically designed for longitudinal data (like Frechet distance or any user-defined distance). Its graphical interface helps the user to choose the appropriate number of clusters when classic criteria are not efficient. It also provides an easy way to export graphical representations of the mean trajectories resulting from the clustering. Finally, it runs the algorithm several times, using various kinds of starting conditions and/or numbers of clusters to be sought, thus sparing the user a lot of manual re-sampling.

Stability and change in positive development during young adulthood

Calls have been made for a greater focus on successful development and how positive functioning can be conceptualized in theory and empirical research. Drawing on a large Australian community sample (N = 890; 61.7% female), this article examines the structure and stability of positive development at two time points during young adulthood. Previously, we developed and empirically tested a model of positive development at 19-20 years comprised of five first-order constructs (civic action and engagement, trust and tolerance of others, trust in authorities and organisations, social competence, and life satisfaction) and a second-order positive development construct. In the current study, we replicated this model at 23-24 years and found that it was again a good fit for the data, and was equally applicable for young men and women. Hence, positive development can be conceptualized according to five important dimensions at both ages. While variable-oriented tests suggested continuity in levels of positive development, person-oriented analyses revealed four distinct patterns of positive development over time-two reflecting stability (stable high [34.5%] and stable very low [11.6%]) and two characterized by change (low/average increasing [30.4%] and average decreasing [23.5%]). There were significant differences in the gender composition of these groups, with young women overrepresented in the more favourable groups. Thus, despite mean level stability, positive development is characterized by change for many young people, suggesting the importance of identifying factors that support young people's capacity for positive functioning over this transitional period. The current findings contribute to our understanding of the nature and course of positive development over this important period.

A taxometric study of hypochondriasis symptoms

Hypochondriasis has been conceptualized as both a distinct category that is characterized by a disabling illness preoccupation and as a continuum of health concerns. Empirical support for one of these theoretical models will clarify inconsistent assessment approaches and study designs that have impeded theory and research. To facilitate progress, taxometric analyses were conducted to determine whether hypochondriasis is best understood as a discrete category, consistent with the DSM, or as a dimensional entity, consistent with prevailing opinion and most self-report measures. Data from a large undergraduate sample that completed 3 hypochondriasis symptom measures were factor analyzed. The 4 factor analytically derived symptom indicators were then used in these taxometric analyses. Consistent with our hypotheses and existing theory, results supported a dimensional structure for hypochondriasis. Implications for the conceptualization of hypochondriasis and directions for future study are discussed.

Latent Class Detection and Class Assignment: A Comparison of the MAXEIG Taxometric Procedure and Factor Mixture Modeling Approaches

Taxometric procedures such as MAXEIG and factor mixture modeling (FMM) are used in latent class clustering, but they have very different sets of strengths and weaknesses. Taxometric procedures, popular in psychiatric and psychopathology applications, do not rely on distributional assumptions. Their sole purpose is to detect the presence of latent classes. The procedures capitalize on the assumption that, due to mean differences between two classes, item covariances within class are smaller than item covariances between the classes. FMM goes beyond class detection and permits the specification of hypothesis-based within-class covariance structures ranging from local independence to multidimensional within-class factor models. In principle, FMM permits the comparison of alternative models using likelihood-based indexes. These advantages come at the price of distributional assumptions. In addition, models are often highly parameterized and susceptible to misspecifications of the within-class covariance structure. Following an illustration with an empirical data set of binary depression items, the MAXEIG procedure and FMM are compared in a simulation study focusing on class detection and the assignment of subjects to the latent classes. FMM generally outperformed MAXEIG in terms of class detection and class assignment. Substantially different class sizes negatively impacted the performance of both approaches, whereas low class separation was much more problematic for MAXEIG than for the FMM.

Seeking verisimilitude in a class: a systematic review of evidence that the criterial clinical symptoms of schizophrenia are taxonic

This review examines whether there is evidence that the criterion symptoms of Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition) (DSM-IV) schizophrenia are taxonic--that schizophrenia is not part of a single distribution of normality. Two taxometric methods, coherent cut kinetics (CCK) and latent variable modeling (LVM), are demonstrated to be sensitive to latent classes and, therefore, were regarded as providing relevant statistical evidence. A systematic literature search identified 24 articles describing analyses of 28 participant cohorts in which CCK or LVM methods were used with one or more criterion symptoms of schizophrenia. Virtually all analyses yielded results that, on first impression, favored taxonic over dimensional interpretations of the latent structure of schizophrenia. However, threats to the internal and external validity of these studies--including biased or inadequate analyses, violation of statistical assumptions, inadequate indicator screening, and the introduction of systematic error through recruitment and sampling--critically undermine this body of work. Uncertainties about the potential effects of perceptual biases, unimodal assessment, and item parceling are also identified, as are limitations in seeking to validate classes with single or double dissociations of outcomes. We conclude that there is no reason to seriously doubt a single-distribution model of schizophrenia because there is no evidence that provides a serious test of this null hypothesis. A second fundamental question remains outstanding: is schizophrenia truly a group of schizophrenias, with taxonic divisions separating its types? We make design and analysis suggestions for future research addressing these questions.

Heterogeneity in antisocial trajectories in youth of adult sexual aggressors of women: an examination of initiation, persistence, escalation, and aggravation

Recent explanatory models of sexual aggression of women have emphasized the role of an antisocial tendency in explaining sexual aggression. If those models agree about the importance of an antisocial propensity, they disagree about the presence of a single or multiple pathways leading to sexual aggression. Currently, no empirical studies have examined within-individual changes of antisocial behavior in youth of sexual aggressors of women and whether those changes are related to the unfolding of the sexual and nonsexual criminal activity in adulthood. This study examines the presence of antisocial trajectories in childhood and adolescence using a sample of 209 convicted adult sexual aggressors of women. A dynamic classification procedure using cluster analyses yields five distinct antisocial trajectories, which are then compared using analysis of covariance on various parameters of criminal activity in adulthood. The results highlight the heterogeneity of antisocial development in youth of adult sexual aggressors of women. Patterns of initiation, persistence, and escalation in youth are related to the general, violent, and sexual offending in adulthood.

A taxometric investigation of DSM-IV major depression in a large outpatient sample: interpretable structural results depend on the mode of assessment

Most taxometric studies of depressive constructs have drawn indicators from self-report instruments that do not bear directly on the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) diagnostic construct of major depressive disorder (MDD). The present study examined the latent structure of MDD using indicator sets constructed from a semistructured clinical interview, self-report questionnaires, and a combination of the two. Taxometric analyses were performed in a large sample of outpatients with primary mood or anxiety disorders. For clinical rating data, results were more consistent with taxonic than dimensional structure, an interpretation supported by additional data obtained from 27 independent raters and objective indices of curve fit. Questionnaire indicators yielded inconclusive results, and combined rating- questionnaire indicators yielded results suggestive of taxonic structure. The findings highlight the importance of assessment in the study of a construct's latent structure and extend recent findings suggesting that MDD may be taxonic.

Stability analysis in K-means clustering

This paper develops a new procedure, called stability analysis, for K-means clustering. Instead of ignoring local optima and only considering the best solution found, this procedure takes advantage of additional information from a K-means cluster analysis. The information from the locally optimal solutions is collected in an object by object co-occurrence matrix. The co-occurrence matrix is clustered and subsequently reordered by a steepest ascent quadratic assignment procedure to aid visual interpretation of the multidimensional cluster structure. Subsequently, measures are developed to determine the overall structure of a data set, the number of clusters and the multidimensional relationships between the clusters.

Assigning cases to groups using taxometric results: an empirical comparison of classification techniques

Determining whether individuals belong to different latent classes (taxa) or vary along one or more latent factors (dimensions) has implications for assessment. For example, no instrument can simultaneously maximize the efficiency of categorical and continuous measurement. Methods such as taxometric analysis can test the relative fit of taxonic and dimensional models, but it is not clear how best to assign individuals to groups using taxometric results. The present study compares the performance of two classification techniques-Bayes' theorem and a base-rate technique--across a wide range of data conditions. The base-rate technique achieves greater classification accuracy and a more even balance between sensitivity and specificity. In addition, the base-rate classification technique is easier to implement than Bayes' theorem and is more versatile in that it can be used when the context of assessment requires that cases be classified despite the absence of latent classes.

A brief taxometrics primer

Taxometric procedures provide an empirical means of determining which psychiatric disorders are typologically distinct from normal behavioral functioning. Although most disorders reflect extremes along continuously distributed behavioral traits, identifying those that are discrete has important implications for accurate diagnosis, effective treatment, early identification of risk, and improved understanding of etiology. This article provides (a) brief descriptions of the conceptual bases of several taxometric procedures, (b) example analyses using simulated data, and (c) strategies for avoiding common pitfalls that are often observed in taxometrics research. To date, most taxometrics studies have appeared in the adult psychopathology literature. It is hoped that this primer will encourage interested readers to extend taxometrics research to child and adolescent populations.

A new method of spatio-temporal topographic mapping by correlation coefficient of K-means cluster

It would be of the utmost interest to map correlated sources in the working human brain by Event-Related Potentials (ERPs). This work is to develop a new method to map correlated neural sources based on the time courses of the scalp ERPs waveforms. The ERP data are classified first by k-means cluster analysis, and then the Correlation Coefficients (CC) between the original data of each electrode channel and the time course of each cluster centroid are calculated and utilized as the mapping variable on the scalp surface. With a normalized 4-concentric-sphere head model with radius 1, the performance of the method is evaluated by simulated data. CC, between simulated four sources (s (1)-s (4)) and the estimated cluster centroids (c (1)-c (4)), and the distances (Ds), between the scalp projection points of the s (1)-s (4) and that of the c (1)-c (4), are utilized as the evaluation indexes. Applied to four sources with two of them partially correlated (with maximum mutual CC = 0.4892), CC (Ds) between s (1)-s (4) and c (1)-c (4) are larger (smaller) than 0.893 (0.108) for noise levels NSR</= 0.2; Applied to four sources with two of them completely correlated, CC (Ds) between s (1)-s (4) and c (1)-c (4) are larger (smaller) than 0.97367 (0.1898) for a random noise level NSR</= 0.2; Applied to 128, 64 and 32 recording electrodes, CC (Ds) between s (1)-s (4) and c (1)-c (4) are larger (smaller) than 0.9557 (0.4251) for a random noise level NSR = 0.15; And applied to the cases of spatially overlapped scalp activities, CC (Ds) between s (1)-s (4) and c (1)-c (4) are larger (smaller) than 0.9083 (0.4329) for a random noise level NSR = 0.15. Finally, the method successfully decomposed the ERPs collected in a spatial selective attention experiment into three clusters located at left, right occipital and frontal. The estimated vectors of the contra-occipital area demonstrate that attention to the stimulus location produces increased amplitude of the P1 and N1 components over the contra-occipital scalp. The estimated vector in the frontal area displays two large processing negativity waves around 100 ms and 250 ms when subjects are attentive, and there is a small negative wave around 140 ms and a P300 when subjects are unattentive. The results of simulations and real Visual Evoked Potentials (VEPs) data demonstrate the validity of the method in mapping correlated sources. This method may be an objective, heuristic and important tool to study the properties of cerebral, neural networks in cognitive and clinical neurosciences.

Taxonic structure of schizotypal personality in nonclinical subjects: Issues of replicability and age consistency

To assess the replicability and age consistency of the taxonic structure and base-rate of schizotypy, 803 university students (21.9 years) and 929 high school students (16.4 years) were administered three self-report measures of schizotypal personality. The two groups came from the same town and were matched on gender. MAXCOV analyses were consistent with a low base-rate taxon of approximately 10% only in the university student group; in the younger group, the three schizotypal personality measures did not show clear evidence of taxonicity. These findings support the hypothesis of the taxonic structure of schizotypal personality in adult subjects, but they raise questions concerning the identification of schizotypy in younger samples.

Applying the Bootstrap to Taxometric Analysis: Generating Empirical Sampling Distributions to Help Interpret Results

Meehl's taxometric method was developed to distinguish categorical and continuous constructs. However, taxometric output can be difficult to interpret because expected results for realistic data conditions and differing procedural implementations have not been derived analytically or studied through rigorous simulations. By applying bootstrap methodology, one can generate empirical sampling distributions of taxometric results using data-based estimates of relevant population parameters. We present iterative algorithms for creating bootstrap samples of taxonic and dimensional comparison data that reproduce important features of the research data with good precision and negligible bias. In a series of studies, we demonstrate the utility of these comparison data as an interpretive aid in taxometric research. Strengths and limitations of the approach are discussed along with directions for future research.

Examining Factor Score Distributions to Determine the Nature of Latent Spaces

Similarities between latent class models with K classes and linear factor models with K - 1 factors are investigated. Specifically, the mathematical equivalence between the covariance structure of the two models is discussed, and a Monte Carlo simulation is performed using generated data that represents both latent factors and latent classes with known amounts of overlap. It is shown that, under certain conditions, the distribution of factor scores can be related to the continuity of the latent space via tests of multimodality as suggested by McDonald (1967) .

A taxometric investigation of the latent structure of social anxiety disorder in outpatients with anxiety and mood disorders

The latent structure of social phobia was examined in a sample of 2,035 outpatients with anxiety and mood disorders to determine whether the disorder operates in a categorical or dimensional fashion. We performed three mathematically distinct taxometric procedures-MAMBAC, MAXEIG, and L-Mode-using five indicators constructed from clinical interview ratings and questionnaire measures of social anxiety symptoms. Results from screening analyses and simulated comparison data consistently indicated that the data were suitable for taxometric analysis. The collective results across procedures, consistency tests, and analysis of simulated comparison data produced converging evidence in support of the conclusion that the latent structure of social phobia is dimensional.

Information-theoretic latent distribution modeling: distinguishing discrete and continuous latent variable models

Distinguishing between discrete and continuous latent variable distributions has become increasingly important in numerous domains of behavioral science. Here, the authors explore an information-theoretic approach to latent distribution modeling, in which the ability of latent distribution models to represent statistical information in observed data is emphasized. The authors conclude that loss of statistical information with a decrease in the number of latent values provides an attractive basis for comparing discrete and continuous latent variable models. Theoretical considerations as well as the results of 2 Monte Carlo simulations indicate that information theory provides a sound basis for modeling latent distributions and distinguishing between discrete and continuous latent variable models in particular.

Profiling local optima in K-means clustering: developing a diagnostic technique

Using the cluster generation procedure proposed by D. Steinley and R. Henson (2005), the author investigated the performance of K-means clustering under the following scenarios: (a) different probabilities of cluster overlap; (b) different types of cluster overlap; (c) varying samples sizes, clusters, and dimensions; (d) different multivariate distributions of clusters; and (e) various multidimensional data structures. The results are evaluated in terms of the Hubert-Arabie adjusted Rand index, and several observations concerning the performance of K-means clustering are made. Finally, the article concludes with the proposal of a diagnostic technique indicating when the partitioning given by a K-means cluster analysis can be trusted. By combining the information from several observable characteristics of the data (number of clusters, number of variables, sample size, etc.) with the prevalence of unique local optima in several thousand implementations of the K-means algorithm, the author provides a method capable of guiding key data-analysis decisions.

Taxonicity of anxiety sensitivity: a multi-national analysis

Taxometric coherent cut kinetic analyses were used to test the latent structure of anxiety sensitivity in samples from North America (Canada and United States of America), France, Mexico, Spain, and The Netherlands (total n = 2741). Anxiety sensitivity was indexed by the 36-item Anxiety Sensitivity Index--Revised (ASI-R; [J. Anxiety Disord. 12(5) (1998) 463]). Four manifest indicators of anxiety sensitivity were constructed using the ASI-R: fear of cardiovascular symptoms, fear of respiratory symptoms, fear of publicly observable anxiety reactions, and fear of mental incapacitation. Results from MAXCOV-HITMAX, internal consistency tests, analyses of simulated Monte Carlo data, and a MAMBAC external consistency test indicated that the latent structure of anxiety sensitivity was taxonic in each of the samples. The estimated base rate of the anxiety sensitivity taxon differed slightly between nations, ranging from 11.5 to 21.5%. In general, the four ASI-R based manifest indicators showed high levels of validity. Results are discussed in relation to the conceptual understanding of anxiety sensitivity, with specific emphasis on theoretical refinement of the construct.

Anxiety sensitivity taxonicity across gender among youth

The present investigation comparatively evaluated the latent class structure and parameters of anxiety sensitivity (AS) among female and male youth using the Childhood Anxiety Sensitivity Index. Participants were 4462 adolescents (2189 females) in grades 7-12 (M(age)=15.6 years). Consistent with prediction, taxometric analyses indicated the latent structure of AS was taxonic in both males and females, demonstrating the taxonic latent structure of AS is similarly observed across gender. Also consistent with prediction, the base rate of the AS taxon differed between genders -- higher for females (12%) compared to males (7%). These findings are discussed in terms of their implications for the study of AS and panic vulnerability among youth.

Taxometric analyses in psychological research and in classification of mental disorders

The issue of whether manifest psychological variables comprise a latent taxonic or dimensional structure is essential to the contemporary psychology and psychiatry. Fortunately, this issue can be adequately addressed by using several taxometric procedures, all developed over the last few decades. This paper describes five of the most commonly used analyses, as follows: Mean-Above-Mean-Below-A-Cut (MAMBAC), Maximum Covariance (MAXCOV), Maximum Eigen-Value, (MAXEIG), Latent Mode (L-Mode), and Maximum Slope (MAXSLOPE). In addition, this paper briefly outlines Paul Meehl?s diathesis-stress model of schizophrenia and highlights the practical importance of taxometric procedures in both contemporary psychology and in the classification of psychiatric disorders.

Evaluating the latent structure of the MMPI-2 F(p) scale in a forensic sample: A taxometric analysis

P. A. Arbisi and Y. S. Ben-Porath (1995) originally proposed that the Infrequency Psychopathology scale, F(p), be used as the final step in an algorithm to determine the validity of a Minnesota Multiphasic Personality Inventory-2 (MMPI-2) protocol. The current study used taxometric procedures to determine the latent structure of F(p) among examinees with profiles that would necessitate the interpretation of F(p) when using Arbisi and Ben-Porath's proposed algorithm. Participants included a subsample of 289 consecutively referred pretrial forensic examinees adjudicated incompetent to stand trial with high Infrequency (F) scale scores, thereby providing a sample that would be expected to have a high base rate of persons with bona fide psychopathology and persons with incentive to overreport psychopathology. Using MAMBAC and MAXEIG, F(p) produced a taxonic latent structure within the subgroup of examinees who obtained raw scores on F of greater than 17. These results support Arbisi and Ben-Porath's original proposal to use F(p) to identify a distinct subgroup of overreported MMPI-2 protocols within forensic psychiatric examinees with high elevations on F. Implications and suggestions for future research are provided.

Can sample-specific simulations help detect low base-rate taxonicity?

The authors examined the role of the sample-specific simulations (SSS; A. M. Ruscio & J. Ruscio, 2002; J. Ruscio & A. M. Ruscio, 2004) procedure in detecting low base-rate taxa that might otherwise prove elusive. The procedure preserved key distributional characteristics for moderate to high base-rate taxa, but it performed inadequately for low base-rate taxa, with low to moderate nuisance covariance, decreasing indicator validity, and increasing average nuisance covariance. Visual inspection of the curves produced by the SSS procedure, combined with the expectation that the inchworm test should produce a clear peak for low base-rate taxonic samples, led to inappropriate conclusions about the suitability of data. The authors speculate that examination of low base-rate taxa using SSS may require other algorithms for generating comparison distributions, consideration of the shape of the whole curve, and/or relaxation of the expectation that a clear peak will emerge with the inchworm test.

Examining the latent structure of negative symptoms: is there a distinct subtype of negative symptom schizophrenia?

Negative symptoms have emerged as a replicable factor of symptomatology within schizophrenia. Although rating scales provide assessments along dimensions of severity, categorization into a negative symptom subtype is typically conducted. A categorical view of negative symptoms is best reflected in the proposal that enduring, primary negative symptoms, or deficit symptoms, reflect a distinct subtype of schizophrenia . Despite an accumulation of findings that support a categorical conceptualization, the data are also consistent with a dimensional-only model where negative symptom subtypologies simply reflect an extreme on a continuum of severity. Using taxometric statistical methods , the present study examined whether a taxonic, or latent class, model best describes negative symptoms in a sample of 238 schizophrenia patients. In order to obtain more stable estimates of symptoms, ratings on the Scale for the Assessment of Negative Symptoms [Andreasen, N.C., 1982. Negative symptoms in schizophrenia: Definition and reliability. Arch. Gen. Psychiatry 39, 784-788.] were averaged across two assessments over a 6-month period. Two taxometric methods, maximum covariance analysis (MAXCOV) and mean above minus below a cut (MAMBAC) identified a latent class or taxon with a base rate of approximately 28-36%. Members of the negative symptom taxon differed from the nontaxon class in that taxon members were more likely to be male and demonstrated poorer social functioning. Taxon and nontaxon schizophrenia patients did not differ in psychotic or affective symptoms. The findings converge to provide support for a categorical view of negative symptoms. Further research is required to replicate the present taxonic findings and to examine characteristics (including possible etiological factors) associated with this negative symptom taxon.

Taxonicity of anxiety sensitivity: an empirical test among youth

Taxometric coherent cut kinetic analyses were used to test the latent structure of anxiety sensitivity (AS) among 371 youth. Anxiety sensitivity was indexed by the 18-item Childhood Anxiety Sensitivity Index (CASI; Silverman et al., J. Clin. Child Psychol. (1991), 20, 162-168). Two sets of manifest indicators of AS were constructed using the CASI: (1) three item-parcel manifest indicators: disease concerns, unsteady concerns, and mental illness concerns; and (2) nine single-item indicators representing each of these three facets of AS. Results from standard and short-scale MAXCOV procedures, internal consistency tests, analyses of simulated Monte Carlo data, and MAMBAC external consistency tests indicated that the latent structure of anxiety sensitivity among youth was taxonic. Estimated base rate of the observed AS taxon ranged between 13.6 and 16.5%. The present findings are discussed in terms of theoretical implications for the study of AS and vulnerability for anxiety psychopathology.

The Latent Structure of Psychopathy in Youth: A Taxometric Investigation

Using taxometric procedures, the latent structure of psychopathy was investigated in two studies of children and adolescents. Prior studies have identified a taxon (i.e., a natural category) associated with antisocial behavior in adults as well as children and adolescents. However, features of this taxon suggest that it is not psychopathy but rather a broader class consistent with the construct of life course persistent antisocial behavior. Because the only prior study of youth used a non-clinical sample, the base rate of psychopathy may have been too low to reveal a psychopathy taxon, especially against the background of a broader and more prevalent antisocial behavior taxon. Therefore, this investigation sought to increase the likelihood of finding a psychopathy taxon (should one exist) by increasing its expected base rate through inclusion of clinical cases in the samples studied. Results produced evidence for both a broad antisocial behavior problem taxon consistent with past research and a much lower base rate taxon consistent with prevalence expectations for psychopathy. These findings support the existence of latent taxa for both psychopathy and a broader class of antisocial behavior problems. Further taxometric research appears to be warranted, which should use a broader array of indicators, with greater specificity to psychopathy.