Overview

As our field embraces early intervention as a model of health care provision for psychotic illness, it continues to seek both prodromal features that reliably predict risk for subsequent psychosis and yet earlier models of care to delay, diminish or even prevent the expression of such risk1. Currently, a nexus for such studies is clinical high risk for psychosis (CHR-P)2, a state preceding but proximal to first episode psychosis (FEP). While earlier prodromal features are recognised, their overall rate of occurrence in adolescents and young adults among the general population has been typically deemed too high for them to have prospective utility relative to CHR-P3.

Yet earlier prodromal features extend to basic symptoms that include subjective disturbances in thought, affect, motor functioning, bodily sensation, perception and tolerance of stress that may be evident prior to CHR-P4. Emphasis on motor dysfunction here reflects two realities. Firstly, motor dysfunction constitutes a more objective, observable index that is amenable to both lay and automated detection. Secondly, motor dysfunction illustrates an under-appreciated, circular epistemological journey: from a historical literature that clearly indicated motor abnormalities to be intrinsic to psychotic illness, through subsequent decades during which these same motor abnormalities were equated primarily with adverse effects of essentially ubiquitous treatment with antipsychotic drugs, to renaissance of recognition that such motor abnormalities are intrinsic to psychotic illness5,6,7,8. Recent studies on the lifetime trajectory of psychosis now appear to coalesce around developmental processes that link psychological and motor abnormalities from the earliest prodromal features, through FEP, to the subsequent course of psychotic illness; furthermore, they are illuminating the developmental pathobiology that may underlie this trajectory and have implications for early intervention.

Motor dysfunction in its social context as a feature of the psychosis prodrome

Two recent reviews have considered evidence that motor abnormalities may be over-represented in CHR-P subjects9,10 and thus may constitute prodromal features that contribute to prediction of risk for psychosis. Furthermore, a recent study has reported patterns of head movements in the course of social interactions to be altered in CHR-P11. Automated analysis techniques for evaluating motor impairments in their social settings12,13 carry potential for yet earlier detection of risk for psychosis by moving assessment of motor dysfunction away from clinical neurological examination into the day-to-day social context of motor patterns in individuals. Such automated techniques may resolve predictors beyond those discernable by even skilled examiners by adding previously inaccessible information. Nevertheless, this approach and its implications for early intervention would be materially advanced by evidence that motor dysfunction within a familiar social context could be identified via direct visual observation by peers such as family members, friends and work colleagues.

Historical acumen during a naturalistic experiment on the psychosis prodrome

We have previously drawn attention14,15 to striking and neglected historical reports on the early psychosis prodrome, dating back almost 150 years. The first of these16 clearly describes early psychological abnormalities that augur “the accession of mental disorder, hallucinations, illusions, delusions” (see ref. 15). These observations, during what would now be described as the psychosis prodrome, were subsequently confirmed in a naturalistic experiment involving two groups of young adults: those who joined the British army in the course of World War I and those who remained in civilian life17,18 (see ref. 14). It is notable that the observations of military physicians on the psychological prodrome of what was then described as the “embryo psychotic”18 were elaborated to include motor dysfunction as follows:

“The bringing of these cases under early observation was helped considerably by the facilities in detecting premonitory symptoms. If a soldier in the army, who has been previously up to the average standard, is noticed to be unable to stand properly to attention on parade, or does not form fours properly, or in any other way does not seem up to his usual standard, it is almost certain that this will be reported by his N. C. O. [non-commissioned officer] to a superior officer, who will ultimately send him to be examined by the medical officer. … In civil cases, on the other hand, this early detection of symptoms is much more difficult, and patients are much more likely to be able to conceal their symptoms”17.

In presaging contemporary studies by a century, these historical observations on the psychosis prodrome first attest the acumen of physicians in identifying its early psychological features in this setting. More importantly, they also highlight the acumen of lay observers in identifying early motor abnormalities in the course of a familiar social context and bringing these to the attention of physicians. Without further refinement in more typical settings, the overall rate of occurrence of such psychological features in adolescents and young adults among the general population may continue to be too high for them to have prospective utility3. However, inclusion of observable motor abnormalities may add a more accessible and informative index in social contexts with which family members, friends and work colleagues are familiar, without recourse to formal neurological assessment.

Notably, such lay observation in familiar social settings would be cost-effective, non-invasive, scalable and immediately actionable by the observer, in contrast to the expense and technical barriers of automated analysis. In contrast, it should be recognised that the above historical observations lack psychometric rigour and reliance on contemporary observations from untrained lay observers carries inherent risks for false positives in early recognition. Nevertheless, the potential utility of these complementary approaches prompts mechanistic consideration of how far back in the processes leading to the manifestation of psychosis can evidence of motor abnormalities be found.

Motor dysfunction from risk for psychosis to long-term course of psychotic illness

Motor abnormalities across the period from psychosis prodrome to FEP should be seen across a broader continuum that extends both before and after this period. Before this period, it is recognised that attainment of motor milestones is delayed in infants who go on to develop schizophrenia in young adulthood19. Subsequently, neuromotor abnormalities are evident in young children long before they go on to manifest schizophrenia20. Furthermore, neuromotor function is impaired through to adolescence in children who, during adolescence, show psychotic-like experiences21,22 that are associated with increased risk for psychotic illness in young adulthood23.

However, after this period motor abnormalities in psychotic illness across its long-term course post-FEP has remained essentially inaccessible: the contemporary imperative to initiate treatment with antipsychotic drugs as early as possible following FEP has necessitated recourse to the historical record to resolve the natural history of movement disorder over the long-term course of untreated psychosis5,6,7,8. Thus, it is notable that the only contemporary study to have systematically addressed this issue has received so little attention. This involved a rare and critical population of subjects with long-standing schizophrenia who had remained untreated with antipsychotic drugs for decades (mean 35 years) due to a dynamic, predominantly family-oriented approach to the management of psychosis in which physical treatments were avoided. Using typical assessment instruments, these subjects showed prominent involuntary movements that differed little in either quality or quantity from a parallel control group of otherwise similar subjects who had received conventional long-term treatment with antipsychotic drugs across decades24; furthermore, involuntary movements in the subjects with schizophrenia who had remained antipsychotic-naïve across decades appeared considerably more severe than the typically mild involuntary movements evident in schizophrenia at FEP6,24.

These unique findings indicate that movement disorder intrinsic to psychotic illness may reflect a process that does not remain static over its long-term course: severity of involuntary movement disorder may, at least in part, increase as duration of untreated psychosis (DUP) increases. This notion is supported by recent evidence suggesting a subtly active, morbid process across duration of the psychosis prodrome and DUP25,26. Yet a fundamental challenge remains: given the above timeline of sequential neuromotor abnormalities from infancy, through early and later childhood, to CHR-P and FEP, how far back in development are the origins of these processes to be found?

Rooting prodromal motor dysfunction in fetal dysmorphogenesis

That attainment of motor milestones is delayed in infants who develop schizophrenia in young adulthood19 indicates a neuromotor substrate that has been compromised during fetal life. In parallel, evolving insights into genetic risk for schizophrenia continue to implicate disruption of brain development that originates in early fetal life27. Furthermore, such disruption has sequelae that are dynamic and, of particular interest in relation to the social context of motor abnormalities, appear context-dependent: that is, along the neurodevelopmental continuum expression of genetic risk appears to vary across brain regions, stages of development and experiences accrued, to include variations across conventional diagnostic boundaries27,28.

Further recent evidence relates to the genetic, embryological and geometric intimacy between development of the brain and the frontonasal prominences of the face during early fetal life. These findings indicate in psychotic illness dysmorphogenesis of the brain-face developmental complex that extends to disruption in the fundamental process of embryonic breaking of left-right symmetry and associated disruption in normal asymmetries of both the brain and face29,30,31; in particular, these findings implicate processes across gestational weeks 7–14 that appear to involve genes associated with risk for schizophrenia29,31,32. An important element in this developmental framework is dysfunction in cortico-thalamo-striatal-midbrain circuits that involve parallel and overlapping cortico-striatal pathways consisting of limbic, associative and sensorimotor subdivisions21,33,34. Such network dysfunction constitutes a viable basis for the confluence of psychological and motor abnormalities that appears evident across the lifetime trajectory of psychosis from infancy, through childhood, to early prodromal features, FEP and long-term course of psychotic illness. Furthermore, such prospective considerations have implications for early intervention.

Implications for early intervention

Despite continuing emphasis on early intervention models for youth mental health1, the most recent meta-analysis finds no evidence that any active intervention investigated to date has a sustained and robust effect on any of the investigated outcomes for CHR-P when compared to control interventions, including transition from CHR-P to FEP2. These chastening results highlight the need for alternative approaches, three of which relate to recent findings and associated propositions.

Firstly, it should be considered whether even during CHR-P the underlying processes may already be so deep-rooted as to show only residual responsivity to any intervention2,27. While interventions may be more effective during yet earlier phases of the psychosis prodrome, when the underlying processes may be more malleable, the challenge is to identify features pre-CHR-P that are sufficiently reliable and utilitarian. Psychological features of the early psychosis prodrome alone appear to be insufficiently reliable within the general vagaries of adolescents and young adults among the general population3. However, these features may be ‘sharpened’ by parallel consideration of context-dependent motor abnormalities during adolescence, particularly in those social (domestic, educational and subsequent employment-related) contexts with which family members, friends and work colleagues are familiar; coalescence of abnormalities across these two domains of dysfunction may be particularly indicative, subject to the caveats noted above.

Secondly, and more heuristically, future interventions might be applied at the earliest phase of the lifetime trajectory of psychotic illness. Recently, Birnbaum and Weinberger argue that as the start of this trajectory originates in disruption of early fetal brain development “principles of prediction and prophylaxis in the pre- and perinatal and neonatal stages may best comport with the biology of schizophrenia to address the early-stage perturbations that alter the normative neurodevelopmental trajectory”27. Clearly, our field is not yet at a stage in understanding brain dysmorphogenesis in the trajectory to psychosis that might indicate any specific modality for addressing prenatal perturbations. However, were such understanding to evolve, a conceptual model might then be the effective introduction of prenatal folic acid supplementation as a public health measure to reduce risk for neural tube defects35 by impacting the underlying process of brain dysmorphogenesis. Currently, profound challenges exist in generalising prenatal implementation of a nutritional supplement for such defects to prenatal remediation of the complex neurodevelopmental gene-environment interactions that likely contribute to dysmorphogenesis in psychotic illness.

Yet, thirdly, conceptual models for interventions between the prenatal period and the early psychosis prodrome are emerging. For example, integrated public health interventions under current investigation seek to remediate risk factors for developmental perturbations during preconception alone, pregnancy and early childhood alone, or both periods sequentially, across the first two years of life36. Such investigations may reveal interventions that are closer to and of potential for offsetting early developmental pathobiology, including that which appears to underlie psychotic illness.