Only few integrative models of schizophrenia were developed with the explicit aim to explain the pathophysiology and symptomatology using more recent findings from empirical investigations. One early attempt of great influence was the “two-hit hypothesis” which used a genetic vulnerability as the first step and subsequent other pathophysiologic influences (biological, environmental, psychological, or other) as the necessary “second hit” in the pathophysiology of schizophrenia.28
Factors predisposing to the development of schizophrenia and factors precipitating its onset may be distinguished. An integrated model based on sociodevelopmental factors involved in the pathophysiology of psychosis was proposed.29
These approaches were extended in the “three hit model” to include neurodegenerative factors which were thought to be induced or accelerated by the disease onset itself (ie, developmental risk factors, precipitating factors, and neurodegenerative factors30
). These hypotheses have gained much empirical underpinning in recent years and can now be refined in that the pathophysiologic factors involved in each of the different “hits” are beginning to be elucidated as interactions of time variable and partly overlapping factors. The theoretical mechanisms for such multiple pathophysiologic factors interacting on the functions of a certain brain region have been described in an example using the prefrontal-limbic system in schizophrenia by Radulescu.31
The central idea of Howes and Kapur32
is that multiple risk factors for psychoses like frontotemporal dysfunctions, genetic factors, prenatal infections, stress, and drugs may lead to a common final pathway of presynaptic hyperdopaminergic dysfunction. Gene-environment interactions could be integrated in this model via epigenetic regulation of genes of dopamine metabolism. In this model, aberrant salience is thought to be the consequence of the hyperdopaminergic state (although it is unclear how hyperdopaminergia leads to aberrant salience), and this is thought to be the decisive psychological function involved in the pathophysiology of psychotic symptoms. Psychosis is viewed as dopamine-driven aberrant salience filtered through the individual's cognitive and sociocultural schemas. The exact diagnosis within the psychosis spectrum reflects the nature of the pathogenic “hits” on the dopamine system coupled with sociocultural factors leading to dopamine pathways as the common final pathways. The kind of relationship between the hyperdopaminergic synaptic state and ensuing symptoms is still not elucidated. Current research in this area focuses on the validation of a salience assessment scale33,34
and empirical investigations in patients with schizophrenia mainly reduced salience network connectivity,35
a correlation between volume reduction in a brain salience network and the clinical phenomenon of reality distortions in patients with schizophrenia,36
an inverse correlation of salience coding and negative symptoms,37
and a correlation of aberrant salience with the presence of delusions in schizophrenia.38
It remains to be determined whether these factors only operate in the pathway to schizophrenia or also in pathways leading to nonschizophrenic psychotic disorders.
In a similar model, Van Os and Kapur17
developed a model of schizophrenia which emphasizes the interaction of gene and environmental factors and which regards schizophrenia as one aspect of a spectrum of psychotic disorders with gradually different degrees of manifestation of psychopathological symptoms (psychotic symptoms, negative symptoms, cognitive disorder, depression, and mania). Compared with the former model, this second model extends to schizoaffective and bipolar affective disorders, and it also discusses the differential roles of certain genetic factors for the different phenotypic pathways. This model is prototypic of the “spectrum” approaches using in this case, a complex analysis of 5 symptom dimensions to order the multitude of psychotic phenomenology. The advantage is clearly that not only schizophrenia but also other psychotic disorders are being included. The empirical basis for this model is still small.
Although not a completely integrated model, conceptualizing psychotic disorders by dimensions of symptoms instead of using categorical approaches also play a role in the current discussion about the revision of the psychiatric classification systems, especially the DSM-IV of the American Psychiatric Association. The evidence of epidemiological studies shows a continuum of psychosis-like experiences in the population, but some categorical aspects also apply, ie, a “psychotic” population can be distinguished.39
As to the role for classification purposes, Craddock and Owen40
discussed a spectrum model that is based on the observation that several genetic risk factors are shared between different types of psychotic disorders like schizophrenia and bipolar disorders. The degree of severity of the presented symptoms is regarded as a continuum with considerable overlap of symptomatology between different mental disorders. Gene-environment interactions could be essential modifiers and determinants of individual phenotypic expression of psychotic disorders. This leads to the proposal that future classification systems should be based on an assessment of the pattern and degree of pathophysiologic symptom dimensions rather than on categorical definitions. Also, psychiatric classification should more heavily rely on knowledge about the neurobiological foundations of the pathophysiology of psychotic disorders. The genetic data also make the inclusion of some genetic forms of mental retardation or autismus-spectrum disorders into any model of psychotic disorders necessary because there is genetic overlap of schizophrenia with this part of the spectrum of mental disorders.
For integrative models of psychotic disorders, the central question arises, which are the substrates of the actions of genetic, psychological, or environmental factors in the pathophysiology of the symptoms of psychosis. Implicit in the beforementioned models is the assumption that neuronal brain cells or their interactions are this central target. Our group proposed a conception of mental disorders in which brain modules are postulated to be the substrates of the damaging factors.14
“Modular psychiatry” rests on the assumption that the physiologic functions of such modules can be defined and measured, that their disturbances in mental disorders can be detected and quantified, and that it can be shown how such disturbances lead to the signs and symptoms of mental disorders. In modular psychiatry, mental disorders are thus based on empirically studied dysfunctions of neuronal circuitry. Such dysfunctions could be modified by gene-environment interactions and epigenetic regulation of neuronal development, maintenance of synapses, and myelination of long-tract association fibers of the brain. Currently, evidence is accumulating by several ways of investigations like EEG and magnetic resonance tomography that such brain modules exist, that they can be identified and analyzed, and that their interactions and hierarchical organization are altered in people with schizophrenia and other mental disorders like Alzheimer’s disease and attention-deficit/hyperkinetic disorder15,41
Altered modularity will reveal itself as decreased or increased centrality (hubness), altered pathlengths, or altered correlation coefficients between brain areas. These alterations lead to a disturbed hierarchical architecture of the human brain modules, and such changes have been associated with cognitive factors and disease course characteristics in people with schizophrenia41,42
including adolescent adults with childhood-onset schizophrenia.43
The currently available empirical evidence for the modular approach is summarized in .
Empirical Evidence for Disturbed Modularity in Patients With Schizophrenia
The next step would be to determine how such altered brain network architectures lead to psychotic symptoms and whether similar alterations of brain modularity and other network characteristics can also be found in persons with nonschizophrenic psychotic disorders, eg, like in Alzheimer's disease or in cases of encephalitis. Such studies are now feasible since methods are available to use modularity analysis in EEG and magnetic resonance imaging (MRI) data. Modules of the brain could become the bridge between the levels of genetic risk factors, functional and structural brain imaging, brain network analyses, and clinical symptoms. However, currently the pathophysiologic mechanisms by which genetic factors and other somatic factors exert their influence on brain modules—or are influenced by mental disorders—are only beginning to be determined. It is still unclear, which brain modules are the targets in individual cases and how this leads to clinical symptoms. The workplan would thus involve firstly an identification of the disturbed modules, a characterization of the kind of disturbances, and the operationalization of methods like MRI or EEG to detect such disturbances. In further studies, it would then need to be shown that the amelioration of such disturbances is measurable and correlates with significant clinical improvements. Then, the assessment of module disturbances could become a key asset for a “modular psychiatry” based on the objective determination of neurophysiological dysfunctions. In modular networks, the frontal lobe may play a central role in controlling behavior (reviewed by Seitz and coworkers48
A central aspect of modular psychiatry is the communication between different brain areas, which basically relies on synaptic neurotransmission. McGlashan and Hoffman49
provided a seminal neural network model of schizophrenia based on synaptic loss and reduced cortical connectivity, which has considerable attractiveness because it leads to spontaneous network acitivity simulating hallucinations, is well in accordance with some experimental findings and provides a unifying framework with testable hypotheses.
In conclusion, all 4 integrated models are based on complex gene-environment interactions with a range of propsychotic factors being combined in individually different constellations to lead to psychotic disorders. A common theme is the conceptualization of a final pathway leading to the disturbance of neural modules in a yet unknown manner, which is accompanied by or leads to a hyperdopaminergic synaptic state. The strengths of these models are their empirical foundations especially in genetic or neurophysiologic studies. This may hopefully lead to objective and quantifiable analyses of the individual risk factors, their interactions and role in the pathophysiology of psychotic symptoms. While several pathophyiologic risk factors may be shared among all persons affected by psychotic disorders, others may play a role only in individual cases. Modular psychiatry combined with quantitative modeling methods may lead to quantified assessments of the kind, directions, and time-variability of interactions of pathophysiologic factors in individual networks of psychotic pathophysiology explaining not only the current symptomatology but also explaining disease courses and providing prognostic information. This should be helpful not only for the purposes of diagnosis and classification of psychotic disorders but also for individualized treatment approaches. Disadvantages are the yet small evidence base and the complexity of the putative interactions with a multitude of interindividually and probably even time-variant pathophysiologic factors. Currently, there is no empirically validated integrative model of all aspects of psychotic disorders, but modular psychiatry with its clearly operationalized definitions and empirical testability holds promise as a useful basis for further investigations in this research area.