Rethinking Thought Disorder

Abstract

It has been 30 years since Holzman introduced a special issue of the Schizophrenia Bulletin entitled “Thought Disorder in Schizophrenia.” He pointed out in his Editor’s Introduction that in contrast to the explosion of interest at that time in the biological aspects of schizophrenia, there were important areas of study that represented “... relatively neglected aspects of the psychopathology of schizophrenia, namely the varieties of thinking disorders (emphasis added) characteristic of schizophrenic patients and their possible underlying mechanisms.” Perhaps presciently, he ended his introduction by expressing hope that the articles included in that issue would lead to further intensive study of the cognitive (emphasis added) dysfunctions in schizophrenia. There has, indeed, been extensive research conducted in further understanding cognitive dysfunctions in schizophrenia, but considerably less so in understanding thought disorder.

We emphasize in this introduction, as did Holzman¹ 30 years ago, the importance of thought disorder and hope this issue of the Schizophrenia Bulletin will stimulate further interest in what is still a relatively neglected area. Unlike Holzman, however, we suggest that thought disorder may ultimately be best studied as a dimensional, cross-diagnostic phenomenon that will have broad implications for understanding and treating psychopathology.

Rethinking Thought Disorder

Thought is a complex phenomenon, through which our understanding of self, world, and reality is constructed. Connected to our capacity for language, thought is critical to communication and plays an important role in how we understand others and make our own experiences known. Breakdowns in the thought system (eg, how one perceives, interprets, structures, and responds to information) is intrinsically linked to difficulties with psychological and social wellbeing and the ability to function adaptively in the world. At their extreme, these disturbances are the core of psychotic experiences.

Thought disorder (TD) is a multidimensional construct, reflecting peculiarities in thinking, language, and communication. Broadly, TD is defined as any disturbance that affects the form of thinking, including the organization, control, processing, or expression of thoughts. Given the breadth of the construct, TD has been defined and classified in a number of different ways. Different perspectives have placed varying emphasis on features related to the contextual appropriateness of ideas, the way in which they are organized, and the language used to express them.

History of Thought Disorder

Disturbances in thought have been central to the conceptualization of psychosis since Kraepelin² introduced the diagnostic construct, dementia praecox, which categorized psychoses that presented relatively early in life and resulted in progressive deterioration of mental functions. For Kraepelin, this deterioration was evidenced by “derailments” and “incoherence” of thought processes and was observable in the speech of patients.

When Bleuler³ reformulated the dementia praecox construct, he emphasized “splitting” of mental functions as the conceptual core of schizophrenia, describing a process in which the psychological force that holds together facets of the psyche—perception, affect, memory, thought, behavior—breaks down. A primary manifestation of splitting was disturbance of thought, which he inferred through observing the speech of patients. He described subtle phenomenological differences in the form of thought disturbances, describing processes in which “the most important determinant of the associations is lacking the concept of purpose³ (p. 15)” and others in which “associations do not become entirely senseless, but they still appear odd, bizarre, distorted³ (p. 19).” For Bleuler, clinical heterogeneity, and the dimensionality it suggested, was critical to understanding the nature of psychosis. He believed that the accurate measurement and classification of thought disturbances would elucidate the mechanisms underlying psychosis.

With the evolution of our modern diagnostic system, other symptoms displaced the centrality of TD to psychosis. This shift was hastened by publication of the DSM-III, which redefined schizophrenia in terms of observable, polythetic criteria, namely hallucinations and delusions. The complexity and conceptual significance of thought disturbances was further reduced with the revision of schizophrenia in the DSM-IV, referred to as simply “disorganized speech.” In short, TD was reduced by fiat to a subsidiary role in schizophrenia and categorical diagnosis was established as the gold standard.

Contemporary Approach to Diagnosis and Thought Disorder

The movement toward a dimensional approach to diagnosis, the Research Domain Criteria (RDoC)⁴ is consistent with TD as a dimensional, multifaceted set of features that present both transdiagnostically and in psychologically well individuals. Although TD is the central defining feature of psychosis, it does not map neatly onto specific diagnostic categories, but may have greater specificity in terms of etiology, course, and treatment. Disturbances in thinking represent a unique and promising pathway through which to understand the nature of psychotic phenomena as well as their subclinical representations. Exploring this from a dimensional perspective has the potential to contribute to a more comprehensive and integrative model of psychosis, which has implications for both science and clinical applications.

Theories of Thought Disorder

In line with the associationist paradigm that predominated psychology at the turn of the last century, Kraepelin and Bleuler conceptualized TD as a result of disruptions in the interconnectivity of ideas. Over the past century, efforts to explain the nature of TD and its significance in psychosis have been theoretically diverse.

Cognitive-Developmental

Language and thought are abilities that we develop over time and thus, it is not surprising that disturbances in these functions have been explained from a developmental perspective. Piaget ⁵ saw schizophrenia as a result of disruptions in the process by which children acquire knowledge and construct an understanding of the world. He believed normative cognitive development occurred over 4 consecutive stages, namely sensorimotor, preoperational, operational, and concrete operational. For Piaget, in schizophrenia, the establishment of object permanence and differentiation from the mother was disrupted in the preoperational stage, leaving the child unable to move past primary egocentricity. As a result, boundaries between self and other, subjective and objective are never fully established. Further, this affected development in later stages, which Piaget believed was evident in a disrupted ability to categorize, common in those diagnosed with schizophrenia.

Vygotsky’s⁶ view of schizophrenia paralleled that of Piaget, but emphasized the sociocultural context of cognitive development. He saw the development of a child’s cognition as a largely social process, founded in cooperative and collaborative dialogue with caregivers. This was at the core of his theory of psychosis, in which he described “the fragmentation and the breaking of that part of the psyche which is involved in the process of formation of concepts is just as characteristic of schizophrenia as the development of the function of formation of concepts is characteristic of adolescence”⁶ (p. 1063).

Psychodynamic

TD from a psychodynamic perspective (drive theory, ego psychology, object relations, and self-psychology) is largely inaccessible as it represents an unspecified failure of a complex array of mental processes that form representations or “internal working models”⁷ of self and other in the context of affect. Emerging early in an individual’s development, these internal cognitive models affect, among other functions, language patterns.⁸

Examined most frequently in the context of personality disorder, TD from this perspective can be construed as an impairment in a cognitive process that leads to fragmented aspects of the self. That is, TD is a loss of cognitive stability underlying the integrated self, such as a failure to inhibit primary process thinking, resulting in an intrusion of uninhibited drive material (eg, libidinal, aggressive, etc.) Some believe this represents regression to an earlier form of thinking, in which internal representations of the self and other become blurred and fragmented, an idea that could be construed as an interpretation of the earlier cognitive hypotheses.^9,10 Of particular note here is that TD relied on a broad range of behavior observed and interpreted by clinicians.

Social Learning

The origins of TD have also been conceptualized from a social-learning perspective, focused on how interpersonal transactions shape the way in which we perceive and make sense of surroundings, form concepts, process and modulate affect, and construct an understanding of the self and others. When early relationships do not provide the opportunity to establish a shared perspective or present inaccurate or inconsistent feedback about perceptual and emotional experience, the child may not develop an organized, stable representation of the self and world.

Singer and Wynne¹¹ posited the role of familial communication patterns in shaping the development of TD, and psychosis more broadly. Their work was based on a social-transactional hypothesis of TD; as explained by Wynne et al¹² “the fragmentation of experience, the identity diffusion, the disturbed modes of perception and communication, and certain other characteristics of the acute reactive schizophrenic’s personality structure are to a significant extent derived, by processes of internalization, from characteristics of the family social organization. Also internalized are the ways of thinking and of deriving meaning, the points of anxiety, and the irrationality, confusion, and ambiguity that were expressed in the shared mechanisms of the family social organization.”

Their research explored how unusual ways in which parents perceived, interpreted, and reasoned about the world interfered with their ability to establish shared attention and construct mutual meaning with their children. Wynne and Singer posited that these disturbed social interactions compromised the child’s development of cohesive, stable mental representations of the self and world, thereby putting them at increased risk of psychosis. They identified a pattern of communication that reliably predicted the later emergence and severity of psychosis in children, which was characterized as vague, fragmented, and contradictory, features recognized in contemporary measures of TD (reviewed below). Interestingly, subtle hindrances to shared meaning were more psychologically detrimental to the listener than overt disruptions.

Familial Basis of Thought Disorder

The familial basis of TD has been well evidenced, not only from a social-learning perspective, but from a genetic perspective. This has been supported by evidence of an over-represented aggregation of TD within families.^13,14 Familial-high risk studies have documented TD in children and adolescents who go on to develop psychotic disorders.^15–19 Further, TD is common in clinically unaffected relatives of manic and schizophrenic patients, differentiating them from relatives of healthy controls.^20–22 Interestingly, the type and severity of TD has been shown to cluster in families.^22,23

Adoption studies have helped to disentangle the genetic and environmental contributions to TD. Wahlberg and colleagues²⁴ examined the likelihood of TD in adoptees, based on genetic-risk (presence or absence of psychotic disorder in biological mother) and communication deviance (CD) in the adoptive parents. Results suggested a significant interaction effect, with high-risk adoptees in high CD environments exhibiting the greatest TD. Interestingly, high-risk adoptees in low CD families had significantly lower TD, suggesting potential protective effects of environment. Results have also demonstrated the co-familiality of TD independent of shared environmental factors. In a subsequent study, Wahlberg and colleagues²⁵ replicated these results using the TDI, finding that the interaction between high-risk status and CD in the adoptive parents distinctly predicted idiosyncratic thinking in adoptees. Taken together, results from family and adoption studies suggest that both genetic and environmental influences are likely involved in the development of TD.

Classification and Measurement

Bleuler used the sentence, “The mountains which are outlined in the swellings of oxygen are beautiful” as a clear depiction of deviant use of language (ref.¹⁴). While the peculiarity of this statement is largely agreed upon, how to describe and define it is a more complex issue. The term TD reflects a broad construct, encompassing diverse conceptual and nosological perspectives cutting across domains of thought, language, and speech. While the study of TD has a rich and lengthy history, consensus has not been reached on what constitutes disordered TD or how it should be classified and measured. A variety of approaches have been used to elicit and assess thought disturbances, including interviews with rating scales, psychodiagnostic tests, and count-based metrics. With this, questions have been raised as to the degree of intricacy with which to measure TD, specifically related to capturing the complexity of language phenomena vs maximizing inter-rater reliability, minimizing subjectivity, burden of learning, administering, and scoring.

Thought, Language, and Communication

The study of “TD” assumes that thought processes can be accurately inferred from speech. This assumption has been challenged by the view that thinking can occur independently of language, suggesting that they are separate domains.²⁶ Studies of language acquisition have supported this; eg, the ability for conceptual thinking has been found to exist in deaf children prior to the attainment of language. The linguist Chaika and Lambe²⁷ has held that psychotic speech is solely reflective of language deficits and does not imply a disturbance in underlying thought processes. This view is supported by observations of lexical and syntactical errors in schizophrenic patients that are also seen in healthy individuals at a lower level.²⁸ Furthermore, individuals diagnosed with schizophrenia who exhibit delusional and bizarre thinking are often able to articulate these thoughts using clear and conventional language.²⁹ This point raises additional questions about the interplay between thought and language, while introducing further complexities to assessment modalities.

Assessment Methods

Rorschach and Rapaport.

A variety of approaches have been developed to measure TD, including reasoning tasks, clinician rating scales, and self-report measures. Perhaps the most widely used medium has been the Rorschach inkblot test, which has a rich history in the measurement of disordered thinking, has been the basis of several scoring systems, and has, paradoxically, even received strong support for this usage from a major critic of the Rorschach.³⁰ The Rorschach lends itself well to TD assessment, as it is a relatively standardized technique that allows examinees to interpret and structure the task freely, minimizing administration variability while maximizing individual response variability.²¹ Its projective nature offers a glimpse into how an individual perceives, interprets, and responds to ambiguous stimuli.

The use of the Rorschach in assessing thinking disturbances can be traced back to Rorschach himself, who observed a characteristic style of responding reflected in the protocols of those with psychosis. He described a tendency for these individuals to formulate responses based on absurd, narrow details of the blot, ignoring typical determinants of form, color, and shading, and often ascribing idiosyncratic or personal meaning to perceptual features.³¹ Building on Rorschach’s work, Rapaport³² sought to develop a more formal system for classifying thought disturbances, which he referred to as “deviant verbalizations.” He believed that adaptive thinking was a function of the integration of perceptual (eg, blot features) and associative (eg, internal ideas, memories, feelings) processes. The relative pull of one process over the other resulted in either a loss or increase of appropriate “distance” from the task, signifying thought disturbances. A loss of perceptual distance reflected a tendency to see the blot as too real, while an increase in perceptual distance was associated with an overly symbolic view of percepts.

Thought Disorder Index.

Rapaport’s system became the basis for the Delta Index,³³ the first standardized assessment to exclusively measure disordered thinking. The Delta Index included 15 of Rapaport’s 21 scoring categories, each of which was assigned a 4-point level of severity. The Delta Index was innovative in that it captured the multifaceted and continuous nature of TD. For this reason, Johnston and Holzman, who developed the Thought Disorder Index, revised the scale decades later.

The Thought Disorder Index²¹ (TDI) provides a system for identifying, categorizing, and evaluating the severity of disordered thinking as expressed in language. TDI scoring can be based on any verbal sample, including the Wechsler Adult Intelligence Scale (WAIS), but is most commonly derived from verbatim responses to the Rorschach Inkblot Test.²¹ As suggested previously, the Rorschach is believed to elicit greater instances of TD than more structured methods,²¹ such as interviews (eg, the Scale for the Assessment of Thought, Language, and Communication²⁶) and nonprojective tests (eg, WAIS).

The TDI specifies 23 categories of thought disturbances, most of which are based on Rappaport’s original classification. Because TD is understood as existing on a continuum, each category is assigned to a level of severity, ranging from mild to severe (0.25, 0.50, 0.75, and 1.00), mirroring the structure of the Delta Index. The 0.25 level reflects subtle instances of cognitive slippage, which are occasionally observed in healthy individuals, particularly in times of anxiety, stress, and fatigue.²¹ Disturbances at the 0.50 level “convey[s] an impression of loss of mooring, shaky reality contact, emotional overreaction, and distinct oddness”²¹ (p. 490). Significant instability in thinking and perceiving is represented at the 0.75 level, while responses at the 1.0 level indicate a complete break from reality.

In developing the TDI, Johnston and Holzman²¹ distinguished 4 qualitative dimensions of TD based on the conceptual relatedness of the individual categories. These include: associative looseness, in which responses appear to be driven by internal processes instead of demands of the task; combinatory thinking, in which percepts, ideas, or images are joined in an inappropriate, incongruous, or unrealistic manner; disorganized responses, in which a lack of clarity of thought and sense of confusion are displayed; deviant verbalizations, in which word usage is odd, idiosyncratic, or undecipherable.

A subsequent factor analysis yielded 6 discrete factors, many of which overlapped with the original, conceptually derived dimensions.²² These “empiric factors” included: (1) combinatory thinking, (2) idiosyncratic verbalizations, (3) autistic thinking, (4) fluid thinking, (5) absurdity, and (6) confusion.

The TDI is a highly sensitive measure of TD and thus is able to detect subtle disturbances in language that may be overlooked by other methods. This is facilitated by the scoring protocol, as ratings are based on written transcriptions of verbatim samples, which allows for systematic analysis of thought disturbances, in terms of qualitative form and severity.

Bizarre-Idiosyncratic Thinking.

Harrow and Quinlan³⁴ introduced the notion of “bizarre-idiosyncratic thinking” to describe a diverse set of language behaviors associated with formal TD. They defined bizarre-idiosyncratic thinking as: (1) unique to the particular subject; (2) deviant with respect to conventional social norms; (3) frequently hard to understand, or to empathize with, in the context from which the response arose; (4) may appear confused, contradictory, or illogical; (5) may involve sudden or unexpected contrasts; and (6) are usually inappropriate in relation to the task at hand. Initially, Harrow and Quinlan rated these domains of disordered thinking based on responses to the Rorschach,^35,36 but later moved away from this due to the demanding and time-consuming nature of Rorschach administration, opting instead for verbal samples derived from free-verbalization techniques.

Scale for the Assessment of Thought, Language, and Communication.

Alternative approaches to assessing TD have been developed that are not as methodologically rigorous as the TDI. Of these measures, the Scale for the Assessment of Thought, Language, and Communication^26,37,38 (TLC) is perhaps the most widely used clinician-rated assessment of TD in both research and clinical practice. Its development was undertaken as part of the broader objective to establish a standard set of TD subtypes for inclusion in the glossary of the DSM-III. Subtypes were identified and defined based strictly on clinical observation, with no assumptions of underlying etiological mechanisms. This atheoretical approach was assumed deliberately in the service of designing an instrument with high inter-rater reliability and clinical utility.²⁶ The original definitions were piloted in a small sample of patients (n = 44) and subsequently revised to improve clarity. The severity of each item is rated on a 4-or 5-point scale, ranging from “absent” to “severe.” These anchor points are defined quantitatively (eg, speech behavior occurs 5–10 times during interview) and are item specific. In addition to the relative severity of each item, TD subtypes are identified as “more pathological” (eg, poverty of speech, pressure of speech, clanging) or “less pathological” (eg, circumstantiality, perseveration, blocking).³⁹ Thus, items are not equally weighted in the determination of global TD severity.

Several studies have examined the factor structure of the TLC, with mixed results. In her early examination of the TLC, Andreasen³⁷ conducted an exploratory factor analysis of 12 TLC items, which yielded a single “verbosity” factor, on which derailment, illogicality, loss of goal, perseveration, incoherence, and pressure of speech loaded positively and poverty of speech loaded negatively. These results were interpreted as evidence for distinct “positive” (florid) and “negative” (diminished) dimensions of TD. In a subsequent factor analysis of the complete TLC, Andreasen and Grove³⁹ reported 3 distinct domains, fluent disorganization, emptiness, and linguistic control. A comparable 3-factor model was generated from an exploratory factor analysis of 8 TLC items²³ and later replicated through confirmatory factor analysis.⁴⁰ Interestingly, however, the inclusion of all 18 TLC items has typically revealed a more complex factor structure, with 6–7 distinct dimensions.^41–43

Communication Disturbances Index.

The Communication Disturbances Index (CDI) assesses communication disturbance in natural speech in terms of categorization and severity without attempting to separate language from thought. The CDI is based in the assumption that both thought and language are cognitive processes that occur together and, therefore, can be measured as such.^44,45 Unlike some measures of formal TD, the CDI is sensitive to low levels of disturbances, including minor communicative errors. This increases the CDI’s utility in identifying those in premorbid and prodromal phases of schizophrenia.

The theoretical basis for the CDI came from linguistic concepts pioneered in the works Cohesion in English⁴⁶ and Crazy Talk,⁴⁷ which focused on the link between speech and “real-world” referents.^28,44 Prominent in both compositions is the idea of cohesion, which drives successful communication through the effective synthesis of speech and its accurate reference to the outside world.⁴⁸ Additionally, earlier studies on linguistics by Docherty and colleagues^44,49–51 helped to inform the construction of the CDI. Each of these 4 studies looked at speech samples from a schizophrenic sample and assessed variables such as unclear and incompetent references in addition to administering established measures of formal TD. Using the aforementioned theoretical framework and research-based information, 6 categories of communication failure were created as part of the CDI; they are: vague references, confused references, missing information references, ambiguous word meanings, wrong word references, and grammatical unclarities.^44,52

Thought Disorder Measurement: Dimensional Vs Categorical.

It is important to distinguish between the qualitative nature of the various scales/ratings and how those scales/ratings are used in data analyses. All of the measures reviewed above are dimensional in nature. Whether based on a clinician’s observation or the counting of instances of a form of TD, numbers are assigned that represent ordinal differences among levels of TD. A “1” reflects less TD than a “3” which in turn reflects less TD than a “5.” That is, the numbers in all measurement systems capture a dimensional severity scaling. The data used to generate the numbers can vary widely, for example, from global ratings of alogia to the number of peculiar instances of word usage. It is important, therefore, to remember that while all measures provide dimensional numerical indices of TD severity, the underlying behavioral data upon which those numbers are based can be vastly different. Finally, it is often the case that dimensional scores will be used to create categorical distinctions. It is not uncommon to find that schizophrenia patients are divided into those “with” and “without” TD for purposes of statistical analysis. This can lend a deceptive sense of differentness between groups that belies the continuous nature of the TD measurement and the “fuzziness” of boundaries among groups.⁵³

Diagnostic Specificity of Thought Disorder

For much of the 20th century, TD was widely accepted as a schizophrenia-specific feature and studied within this context. This perspective shifted in the 1970s, with recognition that thought disturbances were also prevalent in mania, spurring considerable efforts to identify patterns of TD that differentiated among diagnostic groups. The earliest of these studies, which assessed TD using abstract reasoning tasks (eg, Gorham’s Proverbs Test), demonstrated that under-inclusive thinking tended to characterize schizophrenia, while over-inclusive thinking was more typical of mania.^54,55 Using this approach, depressed patients have been shown to exhibit deficits characterized by under-inclusive and concrete thinking compared to controls, although to a lesser degree than schizophrenia patients.^56,57

Similar patterns were found in a series of later studies using the TLC, with poverty of speech and speech content (ie, negative TD) common to schizophrenia, and pressured speech, tangentiality, incoherence, illogicality, and loss of goal characterizing mania (ie, positive TD).^{37,39,49,58,59} However, inconsistent evidence has also been reported, with several studies finding greater tangentiality, looseness, and illogicality in schizophrenia compared to mania.^41,60,61 Schizoaffective patients were included in only 2 studies, both of which evidenced TD profiles comparable to manic patients.^39,61 Examining differences by schizophrenia subtype, Andreasen and Grove³⁹ reported a significantly higher prevalence of poverty of speech content in patients with hebephrenic vs paranoid schizophrenia.³⁹

The presence of negative TD has also been observed in patients with primary depression, although overall TLC severity tends to be considerably lower than that in schizophrenia and mania.^37,61,62 Further, patients diagnosed with psychotic depression tend to exhibit significantly greater alogia, poverty of content, blocking, and perseveration than their nonpsychotic counterparts.⁶²

A series of studies by Holzman and colleagues examined the form and severity of TD in schizophrenia, mania, and schizoaffective disorder using the TDI.^21,22,63,64 Results revealed disorder-specific patterns of TD, in which schizophrenia was characterized by “fluid thinking, interpenetrations of one idea by another, unstable verbal referents, and overly concise and contracted communications which give the impression of inner turmoil and confusion”⁶³ (p. 369). Conversely, manic thought processes were described as “loosely tied together ideas that are excessively and immoderately combined and elaborated” with “a playful, mirthful, and breezy quality to their productions”⁶³ (p. 369). Contrary to the findings of Andreasen and Grove,³⁹ Johnston and Holzman²¹ reported no significant differences in TDI scores between paranoid and nonparanoid patients.

The pattern of TD in schizoaffective patients was less consistent, as significant differences were found between those in manic and depressed states. Schizoaffective manic patients exhibited a high level of combinatory thinking comparable to manic patients. However, they lacked the characteristic flippancy and humor of the manic patients and resembled the schizophrenic patients in terms of the frequency of idiosyncratic verbalizations, confusion, and autistic logic. The schizoaffective-depressed patients were similar to the schizophrenic group in terms of frequency of absurd responses and relatively constricted protocol length. However, they had very low levels of total TD, with their overall TDI scores resembling those of a healthy comparison group. In contrast, their rate of absurd responses was similar to that of the schizophrenic group.

Despite evidence of differences in qualitative form and severity across disorders, the variability in TD presentation is not fully explained by diagnostic groupings. Comparative studies of TD consistently report significant within-group variance (ie, large standard deviations) in TDI total and factor scores, across diagnostic and control groups. This point was empirically demonstrated using discriminant-function analysis, which found that TDI total scores correctly classified only 63.0% of manic and schizophrenic patients.⁶⁴ Qualitative factors were shown to more accurately differentiate groups, as a subset of 5 empirically derived categories (irrelevant intrusions, combinatory thinking, fluid thinking, confusion, and idiosyncratic verbalization) correctly classified 76.5% of the sample. However, when re-examined in a subsequent study that also included a schizoaffective subsample, the accuracy rate of the same factors dropped to 57.7%.²²

Course and Chronicity.

The course of TD is highly variable across individuals. Diagnosis has been shown to account for a portion of this variance, with TD in schizophrenia being more stable and persistent than in schizoaffective disorder and affective psychoses.^65–67 TD in mania has been described as reversible,³⁹ given the observation that disturbances are typically severe during acute phases of illness (eg, hospital admission) and remit completely following treatment.^39,49,66,68 However, remission of positive TD has also been observed in schizophrenic patients,³⁹ suggesting that some subtypes tend to follow an episodic course, independent of diagnosis. Forms of TD that are characteristic of schizophrenia have been shown to have greater stability over time, including more severe level disturbances (ie, 0.50, 0.75) on the TDI^17,67 and idiosyncratic verbalizations.^17,18,57,68

Summary.

Research has documented differences in the form and severity of TD across diagnostic groups. Broadly, thinking in schizophrenia has been characterized by odd, impoverished, and internally driven speech. Classically “schizophrenic” TD tends to have greater stability over time, with increases in severity during acute phases of illness. The disconnected and elaborated thinking disturbances seen in mania tend to be episodic, emerging during acute phases and after remitting partially or completely with treatment. In contrast, measures of thought disorganization (eg, vagueness, perseveration, inappropriate distance, confusion, looseness, fluidity, absurd responses, and incoherence) appear to reflect a more stable trait. It has been posited that the stable thought disturbances associated with schizophrenia are pathognomonic of underlying pathophysiology, while state-related forms are secondary to clinical and situational factors.^14,25

Despite the high degree of within-subject variability across studies, these data demonstrate a clear association between TD and diagnosis. We would note, however, that the association between TD and diagnosis is heavily influenced by the course and/or reversibility of the thought disturbance. That a type of TD may remit does not mean it is ipso facto TD. To the extent that we rely on the course of a symptom to determine if it should qualify as a symptom reflects a logical error (post hoc ergo propter hoc.)

Given that TD can vary over time, it is possible that the studies of TD reviewed above reflect an association between TD and certain clinical or personal features related to diagnosis, such as symptoms, medication, or social functioning. In this case, diagnosis contains important information about the nature of TD that warrants further study. We review below some of the potential influences on TD.

Implications and Future Directions

Bleuler saw TD as the core defining feature of psychotic phenomena, reflective of the “splitting of the psychic functions” that occurred when, in the process of thinking, one’s ideas and feelings disconnect, becoming fragmented and competing functions. This view was echoed by Meehl,⁶⁹ who called TD the “diagnostic bell ringer” for schizophrenia, which for him, was exemplified by the comment, “naturally I’m growing my father’s hair.” While Meehl cautioned that the presence of a single symptom is inadequate grounds on which to infer a nosology, he saw TD as a “rare exception,” which itself was pathognomonic of schizophrenia.

Interest in TD as the conceptual core of psychosis diminished with the rise of the modern DSM system, shifting focus to more clearly defined and readily observed indicators that would enhance the reliability of psychodiagnosis. The last several decades have seen mounting criticism of the DSM framework, prompting a movement to reform our scientific paradigm. At current, this pendulum appears to be swinging back toward a dimension-based approach to classification and study.

With this shift, revisiting the significance of TD as a core dimension in the study of psychosis is not only warranted, but timely. Theoretical and empirical foundations of TD have supported it as a construct with a high degree of specificity (eg, compared to hallucinations, which are highly diverse in terms of etiology). Thus, research in this area has the potential to elucidate robust etiological links, which, in turn, could inform individualized, effective intervention approaches.