Studying Developmental Psychopathology Related to Psychotic Disorders—Challenges and Paradigms in Human Studies

The neurodevelopmental nature of schizophrenia and associated disorders is a central component of most modern concepts of the pathophysiology of these illnesses^1,2 and is the focus of this issue^3–6 While psychosis typically manifests around adolescence and the current diagnostic criteria then usually only allow a diagnosis around this time, strong evidence shows that manifestation is preceded by prodromal high-risk states, often with an attenuated form of psychopathology then seen in full blown illness, which farther back in time are preceded by less specific prodromal states characterized by social withdrawal, depression, and impaired level of functioning,⁷ including motor function.⁸ It is also clear that people who later manifest with schizophrenia already show subtle broad abnormalities in social function and cognition as far back as the first year of life. Studying this broad array of abnormalities across the early life span is challenging, but also critically important because it is a prerequisite to understand disease mechanisms in their interaction with brain maturation and psychosocial development, an enterprise that far beyond the immediate scientific interest carries the hope of yielding markers useful for stratification and precision medicine in psychiatry.⁹ Arango et al.¹⁰ review this field in their lead article in this issue.

Designing human experimental paradigms to study neurodevelopmental psychopathology is thus a very promising and important but also demanding undertaking. Some of the challenges are more pedestrian and largely pertain to methods, while others are conceptual and yet others, possibly the hardest, financial.

On the methodological side, choosing paradigms to study psychopathology is as challenging in childhood as it is in adults, with the added problem that the evidence base tends to be smaller in children. It is even harder to design paradigms that can be used across development given the rapid expansion of cognitive ability and change in the behavioral repertoire. A sometimes overlooked opportunity in this area is to focus on behavioral phenotypes that are more readily apparent and easily quantifiable across the age span, notably motor function. One useful way to avoid cognitive challenge paradigms is to rely on methods such as resting state activity; some encouraging studies have found that brain activity during a range of cognitive functions can be predicted from that activity in adults.¹¹ However, it is not yet clear whether that is true in children and across development.

On the modality side as well, few methods are truly ideal for application across the age span. Although dedicated research groups and advances in methodology are pushing the window of applicability further and further,¹² high-resolution neuroimaging techniques such as magnetic resonance imaging (MRI) require a degree of cooperation that make them to be are of limited utility in the early age span. Electrophysiology, eye movement studies and behavioral observations are applicable early on but are limited in the degree in which they can resolve neural states. One approach to overcome these difficulties is to try and define sources of electrophysiological activities, or neural correlates of behaviors, in high-resource small-scale multimodal studies (combining, eg, electroencephalography (EEG) with functional magnetic resonance imaging (fMRI) or eye tracking with fMRI during scans) and use that information to draw inferences on brain activity from the more accessible technology alone (using information from a multimodal fMRI–EEG study can be used in subsequent EEG—only experiments to compute activity in certain brain regions). Finally, methodological difficulties in linking neurodevelopmental data sets across the age span must be overcome. Already in structural neuroimaging, analyzing data across the age span poses methodological challenges in, for example, co-registration of neuroimages and their subsequent analysis in a standardized space, which is much harder across development given the rapid changes of brain shape, and these issues multiply when multimodal studies are performed. Fortunately, this is a rapidly advancing field.

More complex, at least in my view, are the conceptual issues surrounding developmental psychopathology. While ultra-high risk states share most of the psychopathology of psychosis but in a briefer or attenuated form, this relationship of psychopathological progression weakens as one moves farther away from manifestation. While concepts such as basic symptoms¹³ provide a framework to study the evolution of some research domains (such as thought disorder) continuously up to psychosis manifestation, the fact remains that in earlier states and certainly in early childhood, patients exhibit a nonspecific syndrome of subtle, but broad social, language, motor, and cognitive functional impairments that do not map on the defining features of schizophrenia psychopathology. It one assumes, as in the concept of heterotypic continuity,¹⁴ an underlying disorder/disease process whose behavioral manifestations change over time, it follows that an explanatory account may require an model situated outside psychopathology and behavior and on the level of neural systems. This poses problems for the research domain criteria project as it is currently designed since it is centered around dimensional psychological constructs that are in turn grouped into higher-level domains of human behavior and functioning.¹⁵ If those are discontinuous across develoment, RDoC does not provide the best framework for analysis. Put another way, while development of psychopathology over time may involve (brain) processes that are on a continuous dimension over stages of development, paradigms built on this assumption will be inadequate if the psychopathology varies qualitatively according to developmental stage. Since the RDoC paradigm begins with behavioral constructs, it runs into problems when known neural alterations show different behavorial consequences in different developmental phases. A well-studied example relevant for schizophrenia is afforded by prefrontal lesions, which have strongly time variable effects on behavior that are related to the trajectory in which prefrontal cortex and its connections mature.¹⁶ An even more relevant example of a process that can be well characterized on the level of brain but whose behavioral effects are strongly variable is synaptic pruning, which affects a range of functional circuits diffently depending on their maturational time course.^17,18

How to make these complex interdependencies tractable? We have previously proposed¹⁹ that a useful way forward in this situation is to directly access neural mechanisms as the primary research dimensions, recognizing that not only their clinical consequences, but also their (neuro-) psychological consequences are highly variable across individuals, developmental stages and contexts. One way to identify such systems is to examine clearly identified genetic²⁰ and environmental²¹ risk factors and their interactions, which can be studied in relative isolation in healthy subjects. To adapt this research program to neurodevelopment, an account of risk and resilience mechanism needs to be supplanted by an understanding of human brain development on the systems level and in particular, the identification of changes that characterize especially vulnerable periods, notably fetal and early childhood and adolescence, which might include changes such as synaptic pruning but also more global systems-level connectivity phenotypes such as the ones captured in connectomics.²² While still formidable, this research program is tractable and progress, as reviewed Arango et al,¹⁰ is being made.

To move a research program like this into the study of causal mechanisms, an important consideration for the establishment of interventions, methodologically rich longitudinal studies in neurodevelopment that ideally begin prenatally are necessary. Those require a long-term commitment from funding bodies and researchers for a time span far exceeding usual project funding. It is hoped that the dramatic rewards that can be had not just in scientific discovery but in potential clinical and societal benefits will bring and keep such cohort studies on the scientific agenda of many countries.