Motor Behavior is Relevant for Understanding Mechanism, Bolstering Prediction, And Improving Treatment: A Transdiagnostic Perspective

The Role of The Motor Domain in Psychosis

Abnormal motor behavior was long considered to be of limited value to understanding psychosis. However, research of the past decades has challenged this view by demonstrating strikingly high frequencies of the various motor signs in patients with schizophrenia across the course of the disorder, including subjects at risk or unmedicated first episode patients.^1–3 Motor abnormalities arise spontaneously (linked to pathophysiology) or following psychopharmacological treatment (as side effect), yet the persistence of motor abnormalities across the course of the disorders and the effect of antipsychotics on pre-existing spontaneous motor abnormalities still require further study.^3–5 In addition, studies employing instrumental motor assessments or specific rating scales have emphasized the utility of motor abnormalities for screening the risk for psychosis or staging the disorder.^6,7 Some of the motor phenomena are even able to predict the general outcome of psychosis.⁸ For example, motor coordination deficits in the first episode indicate poor outcome after 10 years.⁹ Critically, hypokinetic motor abnormalities are also associated with low physical activity,¹⁰ which is a key determinant of poor physical health and increased mortality.^3,4 Some motor problems in psychosis extend to other domains of psychopathology, as hand gesture deficits critically impair social interaction.^11,12 Even though motor abnormalities such as tardive dyskinesia or catatonia have been discussed as distinct categorical syndromes, current research efforts have emphasized a dimensional perspective with the possibility of having multiple concurrent motor abnormalities.

Motor signs have been beneficial in shedding light on mechanisms, establishing the neurodevelopmental model of schizophrenia, or informing concepts such as the role of motor circuits and networks in psychosis. For example, neuroimaging studies in psychosis have identified neural correlates of aberrant motor behavior, including white matter tracts relevant to neurological soft signs¹³ or less flexible neural network dynamics in catatonia.¹⁴ Furthermore, using resting-state functional MRI, researchers demonstrated the neural signatures of clinical high-risk (CHR) participants who developed psychosis within one year vs. those high-risk youth who did not; indicating aberrant hyperconnectivity from the cerebellum to cortical motor areas.¹⁵ Unraveling the mechanisms of aberrant motor behavior may also help making progress in informing clinical decisions.

The developments in the field of motor behavior in psychosis also hold promise for improving and informing precision psychiatry efforts. The key questions for precision psychiatry are 1) who among CHR subjects is going to develop mental disorders?, 2) who will have favorable outcomes in the light of an existing mental disorder?, and 3) who will benefit most from specific treatments?. The large cohorts aiming at improving diagnostic specificity have not yet included motor variables into their predication models.¹⁶ Still, motor coordination has been shown to have potential for machine learning in diagnostic applications.¹⁷ Likewise, the disintegration of resting-state sensorimotor networks might become an imaging marker of psychosis.¹⁸ Furthermore, motor abnormalities themselves may become novel treatment targets in psychosis, for example when applying noninvasive brain stimulation or specific motor training.¹⁹

Two major developments have facilitated research in the field of motor behavior in the past decade: First, the decision to separate catatonia from schizophrenia in DSM-5²⁰ and second, the addition of a motor domain to the NIMH Research Domain Criteria (RDoC) initiative.^{21, 22} Both paths emphasized the need to study motor behavior across diagnostic entities. As noted, this past decade of motor research in psychosis has facilitated breakthroughs in prediction, etiology, and treatment domains in psychosis. Given the value of understanding motor dysfunction in schizophrenia, there is enormous potential in evaluating these behaviors in other severe mental illnesses (SMIs), which have distinct trajectories, contributors, and treatment needs, but also share many features with psychosis (e.g., neurodevelopmental disruption, brain circuit dysfunction, needs for improved identification/prediction and treatments). Indeed, promising transdiagnostic studies are already pointing to the utility of incorporating movement into the broader dimensional approach to psychopathology. We discuss behavioral, brain, and neurodevelopmental/risk studies, which support a transdiagnostic approach, survey studies focusing on movement abnormalities in different disorders, providing examples of how this emerging work integrates with the sensorimotor RDoC domain, and then highlight an agenda for future treatment applications.

Transdiagnostic Information on The Motor Domain

Behavior

The prevalence of motor behavior abnormalities across severe mental disorders suggests a key role of the motor domain in psychopathology. Similar to observations in schizophrenia, spontaneous motor abnormalities have been reported across disorders such as major depressive disorder, bipolar disorder, autism spectrum disorders, ADHD, or OCD (e.g. neurological soft signs in 50–100%, abnormal involuntary movements in 10–30%).^{23, 24} Likewise, the catatonia syndrome is prevalent in multiple disorders.²⁵ Furthermore, instrumental measures of motor abnormalities such as velocity scaling (i.e. slowing), force variability (i.e. dyskinesia), or actigraphy (i.e. spontaneous motor activity) demonstrate similarities across psychosis risk, schizophrenia, and affective disorders.^{9, 26–32} Moreover, psychomotor slowing in psychosis is strikingly similar to psychomotor retardation in affective disorders.^{33, 34}

Most of the rating scales used in psychosis research to assess motor abnormalities can readily be applied in other SMI, e.g. scales on neurological soft signs, abnormal involuntary movements, or catatonia (see Table 1). However, self-report of motor behavior was long neglected. The novel sensorimotor and activity scale psychosis-risk (SMAP-R) scale, heavily influenced by the new sensorimotor RDoC domain, demonstrated that self-reported dyskinesia and coordination were specifically high in CHR as compared to adolescents with depression or anxiety, however, the latter groups still had increased values compared to controls.³⁵ Different motor behaviors may represent pluripotent markers, reflecting a common shared vulnerability for SMI, they may indicate a more direct mechanism of psychopathology, or they may represent a common phenotype, driven by different pathogenic factors.

Table 1.

Clinical Rating Scales to Assess Motor Behavior

Motor abnormality	Example behavior	Standard clinical rating scalea
Akathisia	Restlessness or fidgetiness	Barnes akathisia rating scale (BARS)36
Catatonia	Multiple behaviors including immobility, posturing, excitement	Bush Francis catatonia rating scale (BFCRS)37
Dyskinesia	Repetitive, involuntary movements in face or limbs	Abnormal involuntary movement scale (AIMS)38
Neurological soft signs	Problems in coordinating or sequencing motor acts	Neurological evaluation scale (NES)39
Parkinsonism	Rigor, tremor, bradykinesia	Unified Parkinson’s Disease rating scale (UPDRS)40
Psychomotor slowing	Slowing of gross and fine motor behaviors, including gait or handwriting	Salpetrière Retardation Rating Scale (SRRS)41

Note:

^amultiple other scales are available.

Brain Imaging

Alterations of the cerebral motor circuitry are common to multiple SMI.²³ Large-scale transdiagnostic resting-state fMRI studies pointed towards dysconnectivity within the somatosensory-motor network and to altered connection with other networks, suggesting concurrent diagnosis-specific, and general features.^{42, 43} Moreover, Northoff and colleagues suggested that similar neurotransmitter systems are involved in aberrant motor behavior across diagnoses.⁴⁴ The circuits described in the RDoC motor domain comprise cortical motor areas, basal ganglia, thalamus, cerebellum, and connecting fiber tracts.²¹ Parts of this circuitry will be involved in any motor action in health and disease. It has allowed to predict brain-behavior associations relevant to mental disorders, e.g. problems with initiation or inhibition of movement, organization of movement, or temporal dynamics.

Risk

While developmental motor delays and motor abnormalities in adolescence had been traditionally associated with more specific risk of psychosis, findings from the ABCD cohort and others have challenged this notion. In fact, Damme et al. found that family history of depression was linked to motor abnormalities during adolescence, which in turn indicated increased depression risk.⁴⁵ While this points to transdiagnostic potential of motor phenomena, it is important to also recognize that this study focused on the more common psychotic-like experiences,⁴⁶ and that the lower prevalence of formal psychosis in the general population may impact a widespread screening application. Another analysis of the ABCD data set by Damme and colleagues found motor developmental delays to be linked to both depression and psychotic-like experiences in adolescence.⁴⁷ Likewise, familial risk for depression and psychosis was associated with motor developmental delays in infancy. However, motor symptoms during adolescence were distinctly associated with depression or psychosis risk, e.g. psychomotor retardation was linked to depression but not to psychosis risk. Furthermore, motor abnormalities are also found in subjects at increased risk for psychosis or bipolar disorders, such as first-degree relatives who present with psychomotor slowing or neurological soft signs, often in intermediate severity between patients and healthy controls.^48–50 In fact, children of parents with psychosis or bipolar disorder display more NSS, more obstetric complications, and less grey matter volume than community controls, pointing to shared neurodevelopmental risk including motor behavior in SMI.⁵¹ Taken together, the findings suggest that familial risk and early motor signs relate to a general risk for mental disorders, suggesting altered neural maturation. During adolescence, motor abnormalities are quite frequent markers of risks for specific mental illness trajectories, within high-risk populations.^{32, 52–54} The shared familial risk and common early developmental pathways may also account for some of the behavioral and physiologic similarities of the motor domain across disorders. The notion of shared risks is particularly relevant considering the genetic overlap^{55, 56} and the lack of specific pathomechanisms in the diagnostic constructs, which are solely based on symptoms and course of psychopathology, often reduced to episodes in adulthood.

These novel transdiagnostic insights call for a revision of our understanding of the impact of motor behavior in psychosis. We will need to recognize shared and distinct motor profiles across various SMI.⁵⁷

Important Insights From Motor Dysfunction Across SMIs

While not intended to be a systematic review, a brief survey on recent advancements in the understanding of motor phenomena in other SMI highlights the promise of this approach. Currently, the contribution of the motor domain to the general presentation of patients is not similarly understood across SMI.

Autism

Autism is a life-long disorder of neurodevelopmental origin with phenotypic heterogeneity including two core domains, i.e. social communication and restricted, repetitive, or unusual sensory-motor behaviors (RRB).⁵⁸ The RRB domain includes stereotypies and multiple other motor abnormalities, e.g. neurological soft signs. Large pediatric cohorts have indicated that abnormal motor behavior is very frequent in children with autism (35%), linked to symptom severity, and more often detected with standardized tests compared to clinician’s reports (13%).⁵⁹ Likewise, abnormal motor skills are highly prevalent in individuals with autism spectrum disorders (87%), linked to poor functioning and more severe symptoms.⁶⁰ Importantly, heterogeneous neurodevelopmental trajectories will also affect the assessment of motor behaviors in adolescents with autism spectrum disorders.

Multiple neuroimaging studies substantiated alterations of the motor circuitry in autism across the age-span and some studies even described mechanisms for specific motor behaviors. Brain imaging in autism has revealed specific structural alterations, such as abnormal grey matter volume development in the precentral gyrus or the link between white matter properties of the U-shaped fibers between M1 and S1, which is associated with fine motor performance in adults.^{61, 62} In addition, resting-state functional connectivity between thalamus and motor cortex is increased in individuals with autism, for example in infants at high risk for autism at 6 months of age⁶³

Resting-state studies suggested that the somatomotor network was central to many aberrant network connectivity findings. The pattern is shared across behavioral phenotypes, but also differs between children with distinct features of social cognition and RBB.⁶⁴ Moreover, dynamic functional connectivity indicated that in children with autism and RBB, the cortical and subcortical motor networks are more often coupled than in controls. Temporal dynamics of this coupling were linked to scores of repetitive motor actions, but not to scores of cognitive features such as circumscribed interests.⁶⁵ Thus, within the repetitive behaviors, motor actions seem to have a specific neural signature. Ultimately, researchers demonstrated that a stereotypy phenotype was associated with synaptic pathology during neurodevelopmental pruning and cortico-striatal hyperconnectivity in a translational model of autism.⁶⁶ Collectively, findings in autism have shed light on the mechanisms involved in abnormal motor control and repetitive motor behaviors. Psychosis research could benefit from a similar focus on specific brain-behavior associations. Today, studies in psychosis often correlate current motor rating scale scores with functional or structural brain markers. However, scanning subjects with a persistent pattern of abnormal motor behavior instead may help elucidating specific pathobiology within RDoC motor domain subconstructs, e.g. movement initiation or action planning and selection.

Bipolar Disorder

Motor behavior is altered in multiple ways during the course of bipolar disorder. Both processing speed and manual motor speed are reduced in bipolar patients and their healthy relatives compared to controls.^{48, 50} Aberrant motor activity is one of two core symptoms in a network analysis in adolescents at risk for bipolar disorder.⁶⁷ Motor activity patterns have been particularly informative in bipolar disorder. Actigraphy studies indicated lower motor activity in patients in community samples,⁶⁸ also allowing machine learning to correctly classify euthymic bipolar disorder and healthy controls using activity patterns.⁶⁹ Actigraphically assessed motor activity is also linked to white matter microstructure of major motor tracts, such as the left corticospinal tract in patients with bipolar depression.⁷⁰ Aberrant motor behavior may also inform treatment, as psychomotor retardation indicated superior treatment response to ECT in a cohort of 670 patients.⁷¹ During depressive episodes, regional homogeneity is reduced in the somatomotor network and correlates with depression severity.⁷² Importantly, motor abnormalities are also present in euthymic states both on a behavioral and a neural level.⁷³ Moreover, as in psychosis some patients with bipolar disorder present with parkinsonian signs,⁷⁴ and 20% of those seem to have a true dopaminergic deficit unrelated to medication, suggesting that bipolar disorder might be a risk factor for idiopathic Parkinson’s disease.⁷⁵ Thus, motor behavior in bipolar disorder may confer information on risk, course, and outcome. Therefore, motor behavior might become important for precision psychiatry in bipolar disorder.

Major Depressive Disorder (MDD)

Motor behavior in major depressive disorder may be captured using the RDoC sensorimotor domain.³⁴ Albeit psychomotor phenomena have been part of the clinical descriptions of the melancholic subtype of depression, research on the brain-behavior association is still in its infancy. For example, psychomotor agitation and retardation have predictive value for clinical decisions, both indicating superior response to ECT or tricyclic antidepressants.^{76, 77} Actigraphy measures agitation or retardation very well.^{26, 78, 79} In fact, based on actigraphy data, machine learning algorithms may classify MDD from healthy subjects with 84% accuracy.⁸⁰ Thus, motor behavior in MDD may become a relevant marker for diagnosis and treatment outcomes. As in other SMI, MDD is associated with abnormal connectivity in the motor circuit. A large consortium found reduced resting-state connectivity in the sensorimotor network compared to healthy controls.⁸¹ In addition, cortico-motor and basal ganglia circuit alterations have been linked to lower physical activity in MDD.^{27, 78} Psychomotor retardation also correlated with lower global brain signal during an fMRI task.⁸² Furthermore, as in psychosis patients with MDD exhibit problems with correct gesture use.⁸³ Treatment components in MDD include efforts raising physical activity; however, studies have not yet capitalized on assessing treatment effects by focusing on the motor domain. Collectively, assessing motor behaviors in MDD bears the potential for selecting specific treatments according to the presence or absence of motor abnormalities.

Obsessive-Compulsive Disorder (OCD)

The motor behavior of OCD patients becomes evident when observing compulsions or rituals. These behaviors are covered by the RDoC motor domain subconstruct habit. Habit formation is an important process often related to negative valence emotions and clinically relevant to OCD but also addiction. Besides habit, OCD may also present with neurological soft signs and numerous other behaviors such as tics.²³ Much effort has been made in describing the neural circuits involved in compulsivity, which include the motor basal ganglia circuitry but also networks of cognitive control and emotion regulation.⁸⁴ In drug naïve OCD, cortico-pallidal hyperconnectivity has been shown including the SMA, while there was hypoconnectivity in the limbic loop.⁸⁵ Furthermore, the pathophysiology of OCD can be studied using data from deep brain stimulation. In sum, motor behavior and pathobiology in OCD may inform on stereotypic behavior in psychosis.

Potential of The Motor Domain in Clinical Decision Making

Diagnosis and Prediction

Currently, motor developmental delays or single motor abnormalities in childhood and adolescence may indicate an increased risk for SMI. In addition, work in psychosis has pointed towards more specific risk prediction using instrumental motor measures.^{7, 8, 52} In affective disorders, machine-learning algorithms have been applied to classify movement patterns from actigraphy recordings. This approach may be readily tested in psychosis research as well. Likewise, psychosis risk calculators should test the addition of motor signs to increase the predictive value. The prediction might even include information on the network connectivity within the sensorimotor network, which is altered in most SMI also indicating poorer course.⁷ In addition, new studies should test the predictive value of assessing and monitoring motor behavior across SMI before and after the administration of psychopharmacological agents. This will also require more training for clinicians in detecting motor behavior. Motor is better recognized with standardized tests or instruments than clinician’s report across autism, psychosis, or affective disorders.^{6, 26, 58} Finally, novel instrumental means might become the standard assessment of motor behavior.⁶

Personalized Treatment

When we aggregate the findings on the motor domain across multiple SMI, we will eventually be able to inform treatment decisions. Motor behavior may point to specific phenotypes that may have a different course or respond differently to treatments. For some disorders, motor specifiers may become relevant, as in autism⁶⁰ or catatonia.²⁰

Motor behavior would also be a relevant treatment target, as noninvasive brain stimulation or specific motor trainings bear potential to alleviate motor abnormalities.⁸⁶ For example, in autism, motor trainings exerted downstream effects on core symptoms beyond pure motor behavior.⁶⁰

Conclusion

Motor research in psychosis has enabled to identify markers of aberrant neurodevelopment, relevant brain circuitry, and behaviors that are potentially useful for predicting outcomes. We are starting to understand the pathobiology of catatonia,^{25, 87–89} but much is still left to unravel, e.g. the network states leading to psychomotor slowing or coordination deficits. Researchers identified motor abnormalities that are potentially useful for predicting the course of psychosis.⁷ And first studies are testing brain stimulation to ameliorate motor abnormalities.^{19, 86, 90, 91} A truly transdiagnostic perspective on motor behavior would have to include many other disorders, e.g. ADHD or substance abuse. Across SMI, researchers should run more studies testing the mechanisms with combined neuroimaging and brain stimulation or pharmacological challenge. Novel techniques such as machine-learning approaches may help to classify patients according to their movement patterns as in affective disorders. Transdiagnostic studies would also allow for testing the shared neurodevelopmental pathways to SMI. Similar lines of research aimed at describing psychopathology across psychiatric disorders, such as HiToPP. However, HiToPP currently neglects motor behavior, as the early RDoC framework did. We believe that incorporating motor behaviors into standard psychiatric assessments will finally enable more precise prediction and clinical guidance for the benefit of our patients.