Classical Schizophrenia: Liddle and the Core of Schizophrenia

Kraepelin’s inspired assertion that dementia praecox is a specific disease entity has seemingly failed the test of time. Is Peter Liddle rescuing the schizophrenia concept?¹ In this commentary, I will first note how the concept deteriorated and became a roadblock to the acquisition of knowledge. Secondly, I will comment on implications of a clinical syndrome status and Liddle’s proposal in this context.

Describing the course of the schizophrenia concept over time must be brief in a commentary and assertions are for clarity and open to challenge.

1. Kraepelin² specified dissociative thought processes as described by Bleuler, the weakening of the wellsprings of volition leading to emotional dullness and lack of vocation, and a poor prognosis as the three defining features of dementia praecox.

2. Bleuler³ made the case for a specific disease by asserting that the dissociative pathology was primary, fundamental, and present in all cases. He importantly noted that hallucinations and delusions were nonspecific, secondary psychopathology not essential to the concept. In this framework, Bleuler’s “the group of schizophrenias” can be viewed as noting differences among cases in secondary manifestations.

3. In the middle of the 20th century, psychopathology central to schizophrenia was redefined by Schneider.⁴ While there may be a deeper phenomenological basis related to self-disturbances or agency, the Symptoms of First Rank were reality distortion phenomena quite different from the dissociative thought and avolition of Kraepelin and Bleuler.

4. At this time, the field became concerned that the definition of schizophrenia and the application of the diagnosis was not consistent across locations or cultures or individual diagnosticians. Most influential was the report of a substantially broader concept of schizophrenia in the New York City area than in London.⁵

5. In 1981, this issue was addressed by Diagnostic and Statistical Manual of Mental Disorders Third Edition (DSM-III) by giving primacy to Schneiderian First Rank Symptoms (FRS) in the concept and diagnosis of schizophrenia. A single FRS was enough to meet the A Criteria for the diagnosis, and negative symptoms were not even included in the symptom criteria. This despite the only empirical data testing Schneiderian FRS revealed their presence in other psychotic disorders.^6,7 And dividing patients with a schizophrenia diagnosis into a group with and a group without FRS failed to find a difference either in developmental pathology or in the 5-year course refuting the poor prognosis component.⁸ These data were presented to the DSM-III work group as the only available empirical investigation of FRS. Ideology trumped science and FRS helped establish schizophrenia as a specific disease entity highlighting FRS and omitting negative symptoms. Note the shift from dissociative pathology/avolition to reality distortion pathology as critical to the schizophrenia concept.⁹

6. The effect over the next three decades was to promulgate the view of schizophrenia as a disease entity with relative homogeneity in psychopathology and distinct from other forms of psychotic disorders. Virtually, all research was done without addressing heterogeneity within schizophrenia.

7. There were exceptions. Based on the abovementioned FRS studies and an evaluation of the diverse psychopathologies observed in persons with a diagnosis of schizophrenia, we proposed schizophrenia as a clinical syndrome, identified six separable psychopathologies, and proposed positive and negative symptom terminology to address some major differences.¹⁰

8. While slow in overall effect, the second decade of the 21st century has challenged schizophrenia as a disease. Diagnostic and Statistical Manual of Mental Disorders Fifth Edition (DSM-5) is now explicit regarding heterogeneity and clinical syndrome status.¹¹ National Institute of Mental Health announced the Research Domains Criteria paradigm inviting investigation of specific aspects of psychopathology that may cross diagnostic boundaries and may be on a behavioral continuum with the general population.¹² Hierarchical Taxonomy of Psychopathology is applying statistical techniques to reorganize psychopathology into dimensions with groups of pathology joined in different spectrum based on statistical relations.¹³ And genome-wide association studies have a weak phenotype based on a diagnosis of schizophrenia.

Against this backdrop, I consider Liddle’s work in the clinical syndrome construct where two or more specific disease entities may exist. Here Liddle hypothesizes that Kraepelin had it right and that persons with schizophrenia who meet the thought disorder/avolitional/prognosis criteria can be separated from schizophrenia without all the required pathology. He has then shown that these components are defining features that share the same neural network pathophysiology, and this neural network would distinguish this “classical” schizophrenia from other schizophrenia cases within the syndrome framework.

Studies comparing a cohort of persons with schizophrenia diagnoses with a control group have numerous robust confounds that compete with the hypothesis to explain between group differences, not least of all exposure to antipsychotic medications. The within syndrome framework permits testing difference between “classical” schizophrenia and nonclassical schizophrenia where many confounds will be similar between groups. Here the Liddle hypothesis regarding neural network pathology in classical schizophrenia can be tested and results either supporting or falsifying the hypothesis. We previously used this method to test whether deficit schizophrenia (the term we used to indicate a subgroup with primary negative symptoms^14,15) was distinguished from nondeficit schizophrenia. One model called for dorsolateral prefrontal cortex pathology associated with negative symptoms to cause a secondary disinhibition of mesolimbic circuits leading to positive psychotic symptoms. The alternative hypothesis was that the prefrontal pathology would be related to the subgroup with primary negative symptoms distinguishing two groups within schizophrenia, both of whom share mesolimbic system pathology associated with positive psychotic symptoms. As an illustration of strong inference testing, the results falsified the unitary hypothesis model and supported deficit schizophrenia as a separable disease within the syndrome.¹⁶

The search for a disease entity within the schizophrenia syndrome was fruitful with substantial support for the hypothesis.^14,15 In the end, a fundamental problem proved an impediment, something Liddle may avoid. In our work, the issue we could not resolve was whether data supporting deficit schizophrenia as a distinct entity were merely based on the presence or absence of primary negative symptoms. Three concepts are relevant: (1) Schizophrenia is a disease entity with heterogeneity; (2) Schizophrenia is a clinical syndrome comprising two or more distinct diseases (e.g., similar to dementia as a syndrome comprising several separable diseases); or (3) Schizophrenia is a clinical syndrome where specific psychopathology components are separable and explored each in it’s own right.

Put simply, in our work using negative symptoms to subdivide schizophrenia into two groups the possibility that differences on validating measures (e.g., imaging, treatment response, genetics) might be explained by negative symptom pathology rather than negative symptoms as a diagnostic marker of a separable disease entity.

Liddle has contributed extensively to our understanding of brain anatomy and neural system function related to specific psychopathology. He now puts forward a synthesizing and compelling hypothesis. Liddle now marshals evidence for shared neural circuit dysfunction in relation to both disorganizations and impoverishment of mental activities associated with persisting impairment. He further relates these to cognition impairment and proposes validity for Kraepelin’s concept where “classical” schizophrenia separates from other psychoses including those who presently have a diagnosis of schizophrenia. This is a wonderful distillation of decades advancing knowledge in relation to brain mechanisms and schizophrenia psychopathology. It is a perfect setup for strong inference testing as described above. The essential comparative group will be schizophrenia without the critical features. Liddle provides a neural network hypothesis for separating classical from other schizophrenia. In comparing classical and nonclassical schizophrenia, it will be important to consider differences as potentially related to expressive emotional psychopathology in the nonclassical group.

Liddle does not embrace the current trend toward the extended psychosis phenotype replacing schizophrenia and is conceptually opposed to schizophrenia as a reality distortion disorder. Rather, key variables are defined to revisit our history and determine whether Kraepelin and Bleuler can be validated with critical application psychopathology concepts and whether aspects of the classical symptoms cross diagnostic boundaries.

As a final comment, consider the implications for novel therapeutic discovery. To date, actually treatments for some specific attributes associated, but not unique, to schizophrenia are developed for schizophrenia. If classical schizophrenia is based on a shared neural network suboptimal function accounting for thought, avolition, and cognitive psychopathology, it raises the opportunity for treatment discovery with far broader efficacy than found in current therapeutics.