Sensorimotor and Activity Psychosis-Risk (SMAP-R) Scale: An Exploration of Scale Structure With Replication and Validation

Abstract

Background

Sensorimotor abnormalities precede and predict the onset of psychosis. Despite the practical utility of sensorimotor abnormalities for early identification, prediction, and individualized medicine applications, there is currently no dedicated self-report instrument designed to capture these important behaviors. The current study assessed and validated a questionnaire designed for use in individuals at clinical high-risk for psychosis (CHR).

Methods

The current study included both exploratory (n = 3009) and validation (n = 439) analytic datasets—that included individuals identified as meeting criteria for a CHR syndrome (n = 84)—who completed the novel Sensorimotor Abnormalities and Psychosis-Risk (SMAP-R) Scale, clinical interviews and a finger-tapping task. The structure of the scale and reliability of items were consistent across 2 analytic datasets. The resulting scales were assessed for discriminant validity across CHR, community sample non-psychiatric volunteer, and clinical groups.

Results

The scale showed a consistent structure across 2 analytic datasets subscale structure. The resultant subscale structure was consistent with conceptual models of sensorimotor pathology in psychosis (coordination and dyskinesia) in both the exploratory and the validation analytic dataset. Further, these subscales showed discriminant, predictive, and convergent validity. The sensorimotor abnormality scales discriminated CHR from community sample non-psychiatric controls and clinical samples. Finally, these subscales predicted to risk calculator scores and showed convergent validity with sensorimotor performance on a finger-tapping task.

Conclusion

The SMAP-R scale demonstrated good internal, discriminant, predictive, and convergent validity, and subscales mapped on to conceptually relevant sensorimotor circuits. Features of the scale may facilitate widespread incorporation of sensorimotor screening into psychosis-risk research and practice.

Introduction

Several signs of sensorimotor dysfunction have been found to precede and predict the onset of psychosis.^1–8 These sensorimotor abnormalities are tied to cognitive deficits, functional outcomes, and other core features of psychotic disorders.^9–14 A growing body of structural,^15–18 connective, and functional imaging results,^19–21 also has found that sensorimotor features reflect important vulnerability and disease driving mechanism.^11–13,22 Relatedly, sensorimotor abnormalities may be used to form distinct subtypes among individuals meeting criteria for a clinical high-risk (CHR) syndrome.¹⁹ Further, these features may also be useful in monitoring side effects,²³ promoting treatment planning,²⁴ and tracking treatment outcome.²⁵ In addition to sensorimotor abnormalities, sedentary behavior and level of aerobic activity are also important risk/protective indicators for individuals at CHR.^26–29 Despite the clear importance of sensorimotor abnormalities and activity, at the present time, there is no dedicated self-report measure for assessing these symptoms in psychosis-risk populations. Clinician rated scales, designed for other clinical populations (eg, slowing in depression) and purposes (eg, medication side-effect scales) can be applied, but this approach requires highly specialized training. Instrumental methods are sensitive, but typically involve fragile, cumbersome, and/or expensive equipment.^25,30–35 As a result, current options are impractical for widespread clinical use. Additionally, extant self-report and clinical sensorimotor assessments tend to assess a narrow range of sensorimotor signs, but psychotic disorders are characterized by a number of conceptually distinct sensorimotor abnormalities.^35,36

Research has suggested that sensorimotor abnormalities reflect an early vulnerability for psychosis. For example, studies of infants that later develop psychosis in adulthood found that there are elevated levels of coordination deficits, delays, and dyskinesias when compared with controls.³ In addition, a growing body of evidence has indicated that adolescents and young adults meeting criteria for a CHR syndrome exhibit several mechanistically distinct sensorimotor signs. Indeed, coordination deficits and delays in sensorimotor learning may be reflected in cerebellar circuit dysfunction,^17,21,37 whereas slowing as well as hyperkinetic movements related to basal ganglia circuit pathology.¹⁵ Notably, these circuits are relevant to prominent conceptual theories of psychosis,^22,38–41 and relatedly, both domains predicted worsening course,^1,2,21 poor functional outcomes,^42,43 and ultimately conversion to psychosis.^1,36,44–46 Further, neuroimaging studies in this population have refined our understanding of neural underpinnings, as well as the links with disease-relevant mechanisms.^{9,15,22,47–58} With respect to physical activity, those meeting criteria for a CHR syndrome exhibit decreased levels of physical activity.^26,28,29,59 Furthermore, sedentary activity has been linked to abnormalities in the hippocampus among other regions^26,60–62 and relatedly, one open-label exercise intervention has shown evidence of increased cognition, reduced symptoms, and improved functional connectivity in the hippocampus in this population.²⁷

Due to the predictive quality, conceptual relevance, and mechanistic ties, sensorimotor abnormalities and physical activity may be particularly useful to assess in those with a CHR syndrome. Despite its promise, sensorimotor abnormalities and physical activity have been largely under-utilized in the early assessment of psychosis risk; some assessments lack sensorimotor items altogether^63,64 and others only include one^65,66 or 2 self-report items.^66,67 The current study is intended to offer initial validation of a novel (designed specifically for the CHR syndrome) and convenient (self-report) sensorimotor tool, the Sensorimotor Abnormalities and Psychosis-Risk (SMAP-R) Scale.

Methods and Materials

Participants

All participants were recruited as a part of a large, multisite community sample known as the Multisite Assessment of Psychosis-Risk study.⁶⁸ The primary focus of the study was to evaluate markers of risk for psychosis in a large, representative community sample across multiple study sites. Study sites included the greater catchment areas of Philadelphia (Temple University), Chicago (Northwestern University), and Baltimore (University of Maryland, Baltimore County). Recruitment occurred through various outlets, including ads on various Internet sites (eg, Craigslist, Facebook), student volunteer pools, refer-a-friend links, and flyers. Recruitment centered on non-clinical sources; therefore, no recruitment took place at clinical locations such as outpatient psychiatric clinics and hospitals in an attempt to keep the community sample relatively unbiased.

MAP Study—Screening Phase

The study contained 2 phases (figure 1). The first phase was completed by all 3448 participants and included an online battery of measures, including established psychosis risk questionnaires—eg, the prodromal questionnaire⁶³ (PQ), the PRIME Screen⁶⁵—and a variety of other questionnaires, including the SMAP-R Scale. The second phase included subjects that completed the first phase but included a second in-person visit described below. For the current analyses, the presence of a current psychotic diagnosis (n = 3, as assessed during an in-person visit) was also an exclusion criterion. To preserve the representativeness of the community sample, there were no other exclusion criteria.

Fig. 1.

Study sample: guide on data collection phases, sample recruitment overlap, and measure sample sizes.

MAP Study—Validation Phase

The second, validation phase was an in-person visit at the university study site that included a variety of convergent and discriminant measures to validate the current scale. These validation measures were collected within a larger study battery that included clinical assessments using both the Structured Interview for Psychotic Risk⁶⁵ (SIPS) and the Structured Clinical Interview for DSM-5 Research Version⁶⁷ (SCID-5-RV). During the clinical assessment, individuals were classified as clinical high-risk for psychosis (CHR) if they exhibited attenuated positive symptoms or genetic risk and deterioration of function. Attenuated positive symptom criteria were determined by the SIPS guidelines. In total, 85 individuals were classified as meeting criteria for a CHR syndrome, with all other individuals considered community sample non-psychiatric volunteers (CSV). The SCID provided comorbid diagnoses across the entire community sample. All diagnostic decisions were made by trained clinical staff under the direct supervision of each sites’ coauthors VAM, LE, and JS, with weekly cross-site diagnostic consultation calls. The second phase included the completion of the Pennsylvania Computerized Neurocognitive (PNC) battery and additional questionnaires. The PNC battery is a validated, reliable neurocognitive testing platform⁶⁹ that has a demonstrated sensitivity to psychosis symptoms in patients with schizophrenia⁷⁰ and in populations at CHR.³⁰

Sensorimotor Abnormalities Psychosis Risk Scale

The SMAP-R scale is a 14-item questionnaire that included questions about early developmental motor delays, the frequency of abnormal sensorimotor experiences, general assessments of sensorimotor function, and frequency of physical activity. Questions were developed based on a qualitative review of extant literature^{12,22,35,48,71} regarding sensorimotor abnormalities linked to symptoms,^{4,10,19,21,55,72} conversion,^1,5,36,44,45 and clinical/cognitive subtyping¹⁹ in those meeting criteria for a CHR syndrome. From this review of the literature, a set of items was compiled to examine relevant constructs that would also lend well to self-report questions (eg, ballistic movements, sway/balance). These items included several Likert-type scale responses, such as “On a scale of 1–10, how would you rate your current motor skills?” and “Do you enjoy hobbies where you use skilled hand movements (eg, art, crafts)?” (Responses: Never Sometimes Lots). For a version of the final measure and scoring guide, see supplementary materials.

SIPS Risk Calculator

The SIPS-RC is a Structured Interview on Prodromal Risk Symptoms (SIPS)-based risk calculator that provides practical, individualized risk assessments that can be easily implemented in a clinical setting.^73,74 In this calculator, risk probability assessments are derived from 4 dimensions: positive and negative symptom severity, deterioration of function over the past 12-months, and low levels of dysphoric mood. In the current study, the global functioning scale⁷⁵ was used to assess functional decline in the past year. Positive and negative symptom severity was assessed during the SIPS interview and quantified as a composite score of endorsed positive and negative symptoms.⁷³ Finally, low levels of dysphoric mood were assessed during the general symptom assessment of the SIPS.^73,74 This approach has been validated against external independent risk calculators.

Behavioral Validation and Discrimination

The Pennsylvania Neurocognitive Battery was completed at the in-person visit. This battery included a computerized finger tapping (CTAP) performance.⁷⁶ During the CTAP task, participants were instructed to press the spacebar as quickly as possible, moving only the index finger for 10 seconds after their first button press. In this task, trial blocks alternated between the dominant and non-dominant hand for 10 blocks (5 blocks per hand). Median finger tapping and tapping variability (the standard deviation in taps between blocks) behavior was used as a convergent measure of motor performance to assess the sensitivity of the SMAP-R scale to motor performance. Variability in performance was included because a recent study demonstrated that it may be a sensitive and mechanistically relevant indicator among those meeting criteria for a CHR for psychosis.⁷⁷

Analytical Strategy

Exploratory factory analyses were used to examine structure of responses in the 2 analytic datasets (ie, exploratory and validation). These 2 analytic datasets, with independent data-points include 2 analytic datasets: (1) an exploratory, screening analytic dataset (n = 3109) that completed only phase one of the study and (2) a scale validation analytic dataset (n = 439) that completed both phase 1 and phase 2.⁷⁸ For each analysis, all participants with the data for that given analyses were included, rather than reducing the study sample size to only those individuals with a full and complete dataset; the intention in this approach was to both maximize power and transparency, see figure 1 for analyses sample sizes. Factor analyses and item reliability was assessed using the R v.4.0.0⁷⁹ and the Psych package; for more information on how factors were assessed, see supplementary material.⁸⁰

All items that were grouped into a factor were examined as a composite subscale in the subsequent analyses. Items that were not included in the subscales (ie, items that were not related to other items) were treated as independent items and evaluated for discriminant validity and convergent validity. Independent items that failed to show discriminant and convergent validity were excluded from the final scale. These subsequent analyses probed the discriminant and predictive validity of these composite subscales and reduced the total number of comparisons.

In terms of discriminant validity, we examined the degree to which these scale factors differ between relevant CHR group (n = 84) compared other disorders (ie, depression; n = 35 and anxiety; n = 78), and a CSV (n = 78) samples among the validation analytic dataset, using general linear models for the composite scales and chi-square for independent items. In terms of predictive validity, scale factors were related to the calculated risk for psychosis (SIPS-RC) and current symptom severity (SIPS total scales) within the CHR group from the validation analytic dataset. To examine convergent validity, the sensorimotor scales were compared to the motor performance (ie, finger tapping task) in the validation analytic dataset. To examine discriminant validity of behavior, the sensorimotor scales were compared to the performance on an independent attentional task (ie, NL-CPT; see supplementary material for analyses and details).⁸¹ All validation analyses were completed in SPSS version 26.

Results

Participants

There were no significant differences in key demographic variables, including biological sex at birth, ethnicity, race, age, and income across exploratory and validation analytic datasets, as well as discriminatory clinical samples, table 1.

Table 1.

Demographic Metrics: A Comparison of Diagnostic Groups Within the Validation Sample for Discriminatory Analyses

	Exploratory Sample	Validation Sample	Sample Comparison Statistics	CHR	CHV	MDD	Anxiety	Discriminant Sample	Discriminant Group Statistics
	n = 3009	n = 439		n = 84	n = 78	n = 35	n = 78	n = 275
Age -M(SD)	20.31 (2.36)	20.13 (1.84)	t(3436) = 1.523 P = .13
Sex (% Female)	67.75%	77.47%	χ 2(1) = .285 P = .59	75.60%	70.50%	80%	84.60%	74.69%	χ 2 (3) = 4.70, P = .20
Hispanic ethnicity	10.38%	10.25%	χ 2 (1) = .007 P = .93	7.31%	8.97%	14.29%	5.26%	8.98%	χ 2 (3) = 2.76, P = .43
Race			χ 2 (5) = .852, P = .97
American Indian	0.36%	0.46%		0.35%	0.00%	0.00%	0.00%	0%	χ 2 (3) = 2.31, P = .51
Asian	24.06%	25.06%		6.36%	8.13%	2.47%	8.13%	28.98%	χ 2 (3) = 2.45, P = .49
Native Hawaiian	0.03%	0.00%		0.35%	0.35%	0.00%	0.00%	0.82%	χ 2 (3) = 1.417, P = .70
Black/African American	13.51%	14.81%		4.95%	4.59%	1.41%	4.24%	17.55%	χ 2 (3) = .629, P = .89
White	51.34%	50.57%		16.61%	13.43%	7.42%	13.07%	58.37%	χ 2 (3) = 2.59, P = .46
Multiracial	6.16%	6.61%		1.06%	2.12%	1.06%	1.41%	6.53%	χ 2 (3) = 1.78, P = .62
Unknown	7.45%	2.51%		0.71%	1.06%	0.71%	0.00%	2.86%	χ 2 (3) = 3.978, P = .26

Note: CHR, clinical high-risk; CHV, community healthy volunteer; MDD, major depression disorder.

Exploratory Factor Analyses

In the exploratory analytic dataset, the traditional Cattell’s scree analyses suggested that the intercorrelation matrix showed 4 components; scree analyses compared to simulated data suggested the presence of 3 components or 2 factors, supplementary figure 1. Both analytic datasets showed similar scale structures with 2 components dividing the scale into “Sensorimotor Abnormality” subscale items and “Physical Activity” subscale items; when reorganized into 3 factors, the Sensorimotor Abnormality items divided into facets “Coordination Abnormalities” facet items and “Dyskinesia” facet items, table 2. In the validation analyses, the traditional Cattell’s scree analyses suggested that the intercorrelation matrix showed 3 components; scree analyses compared to simulated data suggested the presence of 3 components or 2 factors (figure 2). There were 2 items that did not load onto these factors, including an item on sensorimotor delays (ie, “Are you aware of any delays in walking, talking, or toilet training when you were an infant or toddler?”) and an item on fine sensorimotor hobbies (ie, “Do you enjoy hobbies where you use skilled hand movements (art, crafts)?”). These items were treated as independent items in the following validation analyses. Both exploratory and validation analytic datasets showed similar internal consistency (table 3).

Table 2.

Scale Items Organized by Subscales and Facets

Subscales	Facets	Definition	Items
Sensorimotor abnormalities
	Dyskinesiaa	Involuntary movements most commonly seen in the limbs, mouth, or tongue resembling either chorea (ie, flinging, jerking) or athetosis (ie, writhing)15,43,56
			Do you experience urges to twitch or move suddenly, or make noises?
			Do you ever notice that you have moved suddenly without planning it (eg, leg or arm jerking; head or neck jerking)?
	Coordination abnormalitiesa	Abnormalities in sensory integration, motor coordination, motor sequencing, and reflexes6,82
			Do you feel clumsy?
			Would other people say you are clumsy?
Physical activity		Physical activity, including exercise and general less strenuous activity, is reduced in psychosis and psychosis risk populations26,28
			How often do you participate in physical activities (jogging, team sports) each week?
			On a scale of 1–10, how would you rate your current motor skills?
			How secure/balanced is your current bicycle riding?
			How often do you enjoy participating in sports?
			Do you enjoy aerobic exercise? (jogging, running, swimming)

Note: ^aFor an illustration of these specific types of motor abnormalities, please see: van Harten and Mittal,⁸²Figures 1 and 3.

Fig. 2.

Factor structure for both a 2-factor (Sensorimotor Abnormalities and Physical Activity Subscales) and 3-factor solution (Dyskinesia facets, Coordination Abnormalities facets, and Physical Activity Subscale).

Table 3.

Reliability of Scales Replicated Across Samples

Factor Scales	Exploratory Sample		Validation Sample
	Cronbach’s α	Guttman’s λ	Cronbach’s α	Guttman’s λ
Sensorimotor abnormalities	.69	0.71	.71	0.73
Coordination abnormalities	.71	0.68	.75	0.71
Dyskinesia	.67	0.51	.68	0.51
Physical activity	.68	0.64	.69	0.64

Discriminant Validity Among Clinical Samples

Composite subscales totals were compared across diagnostic groups in independent analyses of variance model (ANOVA) (figure 3). For each of the independent items, a limited set of Likert-responses were compared across groups using a chi-square analysis. Follow-up discriminant function analyses were conducted comparing CHR and CSV groups and showed converging results (supplementary materials).

Fig. 3.

Discriminant validation analyses: composite scales by group.

Sensorimotor Abnormalities Subscale by Diagnostic Group

The Sensorimotor Abnormality subscale significantly differed among groups, F(270,3) = 9.50, P < .001, partial${\eta }^2$ = .10, such that the CHR group (M = 6.72, SD = 3.25, SEM = .36) showed increased Sensorimotor Abnormality score than individuals with depression (M = 5.32, SD = 2.78, SEM = .37, P = .02), anxiety (M = 5.18, SD = 3.20, SEM = .37, P = .006), and the CSV (M = 4.29, SD = 2.21, SEM = .25, P < .001),

Abnormalities Facets by Diagnostic Group

The Coordination Abnormalities facet significantly differed among groups, F(271,3) = 6.193, P < .001, partial${\eta }^2$ = .07, such that the CHR group (M = 5.33, SD = 2.55, SEM = .28) showed increased Coordination Abnormalities facet scores compared to individuals with depression (M = 4.40, SD = 2.44, SEM = .412, P = .05), anxiety (M = 4.51, SD = 2.74, SEM = .317, P = .043), and the CSV (M = 3.70, SD = 1.83, SEM = .21, P < .001). The Dyskinesia facet significantly differed among groups, F(270,3) = 8.54, P < .001, partial${\eta }^2$=.09, such that CHR group (M = 1.41, SD = 1.26, SEM = 0.14) showed an increased Dyskinesia facet score than individuals with depression (M = 0.88, SD = 1.15, SEM = 0.20, P = .02), anxiety (M = 0.67, SD = 0.96, SEM = 0.11, P < .001), and the CSV (M = 0.636, SD = 1.01, SEM = 0.12, P < .001).

Physical Activity Subscale by Diagnostic Group

The Physical Activity subscale significantly differed among groups, F(250,3) = 8.536, P < .001, partial${\eta }^2$=.09, such that CHR (M = 7.47, SD = 2.67, SEM = 0.31) showed less physical activity than the CSV (M = 9.19, SD = 2.74, SEM = 0.32, P < .001). Although the CHR group showed less physical activity than individuals with depression (M = 8.59, SD = 3.18, SEM = 0.48, P = .076) and anxiety (M = 8.19, SD = 3.53, SEM = 0.45, P = .16); this difference was not significant.

Independent Items by Diagnostic Group

The motor delay response (no delays, some delays, a lot of delays) were compared across the diagnostic groups (CHR, CSV, depression, anxiety); groups differed in terms of motor delays χ ² (223) = 9.45, P = .009. The CHR group showed the highest percentage of motor delays with 16.5% of the CHR group reporting some or severe motor delays, and 100% of individuals reporting severe motor delays were also members of the CHR group. The frequency of engagement in motor hobbies (never, sometimes, often) were compared across the diagnostic groups (CHR, CSV, depression, anxiety); groups did not differ in terms of their engagement in fine motor hobbies, χ ² (275) = 8.94, P = .27.

Convergent Validity to Sensorimotor Performance

Composite scales were compared to performance on a finger tapping task performance. Within the CHR group, a repeated-measure general linear model compared motor speed across hands (dominant, non-dominant) by median and variance as the within-subject factor and the SMAP-R subscales, respectively, as between-subject factors. Finger tapping performance related to the “Sensorimotor Abnormalities” subscale, F(71) = 5.83, P = .016; higher self-reported Sensorimotor Abnormalities subscale related to lowered median taps on both the dominant (r_partial = −.23) and non-dominant hands (r_partial = −.18). Finger tapping performance also related to the Coordination Abnormalities facet, F(71) = 5.95, P = .017, such that increased Coordination Abnormalities facet scores were related to lowered median taps on both the dominant (r_partial= -.26) and non-dominant hands (r_partial= −.22). Finger tapping did not significantly relate to any other subscales or items, Ps > .13.

Divergent Validity to Non-Sensorimotor Behavior

Subscales were compared to efficiency on a number-letter continuous performance task. Within the CHR group, a general linear model predicted task efficiency with both SMAP-R subscales as between-subject factors. There were no significant relationships of any subscale, facet, or independent items to attentional task efficiency, Ps > .31.

Predictive and Convergent Validity to Clinical Features

Composite subscales totals significantly correlated with the psychosis risk score (SIPS-RC). Sensorimotor Abnormalities subscale related to SIPS-RC scores, r(312) = .15, P = .008. Coordination Abnormalities facet related to SIPS-RC scores, r(310) = .16, P = .006. Dyskinesia facet related to SIPS-RC scores, r(308) = .12, P = .04. “Physical Activity” subscale related to SIPS-RC scores, r(284) = −.150, P = .01. For each of the independent items a limited set of Likert-responses were compared across groups using a general linear model. SIPS-RC did not relate to infant delays (P = .23) or frequency of fine motor hobbies (P = .63).

Discussion

The SMAP-R scale showed a replicable structure that grouped items into Sensorimotor Abnormalities and Physical Activity subscales. Sensorimotor Abnormalities further divided into Coordination Abnormalities and Dyskinesia facet. Each of these factors is of significant interest to understanding, identifying, predicting, and treating individuals at CHR for psychosis. Therefore, a sensorimotor scale that differentiates distinct sensorimotor categories has significant potential. Further, the SMAPRS subscales discriminated individuals at CHR from anxiety and depression groups and a community sample of CSV. There was modest support for convergent validity. Specifically, the general Sensorimotor Abnormality subscale and the Coordination Abnormalities facet showed a small, but significant relationship to finger tapping performance, but not attentional performance on an unrelated task. Additionally, these scales related to a validated measure of risk for psychosis in the SIPS-RC. Collectively this scale may provide unique insight into sensorimotor abnormalities, which is largely under-utilized in current approaches to researching and treating the psychosis risk syndrome.³⁵ This potential is especially important as sensorimotor abnormalities have demonstrated utility in domains predicting course,^1,2,21 functional outcomes,^42,43 and conversion to psychosis.^1,36,44–46 As a result, the current self-report measure may be used as an initial screening to focus further clinical sensorimotor assessment and targeting individualized treatment.³⁵

The exploratory and validation analytic dataset factor analyses suggested that items of the scale could be grouped into 2 subscales (Sensorimotor Abnormalities and Physical Activity subscales) with 2 facets (Coordination Abnormalities, Dyskinesia) in the exploratory analytic dataset.²² Scale structure was consistent across independent analytic datasets providing converging evidence of the stability and reliability of these subscale structures.⁷⁸ It is also notable that both the exploratory and the validation analytic datasets reflected a diverse community sample, which should increase the generalizability of the current scale structures.⁸³ The 2 subscales identified reflect distinct vulnerabilities/mechanisms/circuits that are central to the etiology of psychosis.²² For example, sedation and physical activity have been related to hippocampal dysfunction, a factor prominent among those identified as meeting criteria for a CHR syndrome,^27,84–86 and are also a central component of prominent conceptual models.⁸⁷ Further, poor coordination,^18,21 as well as issues with balance and ataxia,^10,21,50,88 are relevant to cerebellar-thalamic dysfunction and cognitive dysmetria⁴⁹ and dyskinetic movements and basal ganglia circuit vulnerability are relevant to emerging psychosis as well as the revised dopamine hypothesis.^{9,15,38,53,54,89} Finally, the validation analyses provided initial evidence of the utility of these scales.

The validation analytic dataset was a diverse community sample⁸³ that included CSV as well as individuals with psychiatric diagnoses, including major depression disorder and anxiety disorders, both known to be associated with lower rates of sensorimotor abnormalities (ie, psychomotor slowing/agitation).^83,90 The CHR group showed significantly more Sensorimotor Abnormalities subscale, Coordination Abnormalities, and Dyskinesia facets, suggesting that these scales are particularly sensitive to the presence of psychosis liability, beyond general psychopathology present in depression and anxiety. In contrast, the Physical Activity subscale did not discriminate between the CSV and the CHR group. However, the Physical Activity subscale was sensitive to the presence of general psychopathology as the psychiatric groups did not differ from each other, but differed from the CSV sample.^71,90 It is notable that both the Sensorimotor Abnormalities subscale and the Physical Activity subscale distinguished other anxiety and depression from CSV. This sensitivity to other psychopathology suggests that this scale may also have some utility in examining sensorimotor abnormalities in depression and anxiety. In summary, the scales showed both a specific relationship to psychosis risk and a general sensitivity to discriminate individuals with psychopathology from a CSV group.

Despite the value demonstrated in discriminating groups, the ability to distinguish those meeting criteria for a CHR syndrome from the comparison groups does not directly speak to the SMAPRS’s capacity to measure actual sensorimotor behavior.^91–93 To this end, we have further validated these scales with the finger-tapping task, a measure known to reflect disturbances in the function of underlying neurocircuitry of sensorimotor systems.^30,94,95 This task is particularly probative of cortico-cerebellar interactions that govern the execution of sub-second responses, but also motor slowing more generally. The Sensorimotor Abnormalities subscale related to the performance on this task, driven in large part by the Coordination Abnormalities facet. Although the relationship to finger tapping motor performance was small, it does suggest that individuals in the CHR group were able to faithfully report on their sensorimotor abnormalities (which likely includes other features in addition to motor slowing). Additionally, as expected, none of these subcscales related to performance on a foil (attentional) task.⁸¹ Collectively, these results demonstrated additional discriminant validity as the subscales specifically predicted task performance in the sensorimotor domain and not behavioral performance in unrelated domains.

All of the SMAP-R subscales related to the calculated risk for developing psychosis—a composite score that weights the predictive features of clinical high-risk for psychosis (ie, SIPS-RC score). The calculated risk scale serves as an estimation of the likelihood of an individual to transition to psychosis, but it also takes into account critical features of early clinical risk, eg, symptom severity and decreased global function.^73,74 The relationship of these subscales with calculated risk may reflect the relevance of both Sensorimotor Abnormalities and Physical Activity subscales to emerging psychosis.

The motor developmental delays item distinguished between clinical groups with a larger proportion of individuals in the CHR group who endorsed minor motor developmental delays,^3,96 and endorsements of major motor developmental delays were made up entirely of those in the CHR group. The motor delay item was included in future versions of the SMAP-R scale, but the fine motor hobby item will not be included in the final scale as it did it reflect group membership, clinical risk features, or motor performance.

This study included several strengths, but there were limitations that should be explored in future studies. First, the current study established the items based on a qualitative review, but future studies should consider implementing a search algorithm (eg, PRISMA) with established guidelines to create items. Although the current paper identified 3 scales that may reflect distinct vulnerabilities/mechanisms/circuits, we did not directly assess mechanistic ties and specificity in the current study. Similarly, while the current work assessed internal reliability, we did not assess test-retest reliability within subjects. Future studies will need to assess this critical aspect of reliability. Further, the original version of the scale omitted additional sensorimotor behaviors such as gestures and catatonia that are likely to tap into other distinct vulnerabilities.^{16,88,97–103}

The current study reflects a major advancement in the accessibility and dispersal of sensorimotor assessment, but there is still room for important future work in this area. While the measure contains exploratory gesture items, this questionnaire could be administered with extant gesture self-report questionnaires,¹⁰⁴ which have been validated in psychosis¹⁰⁵ and major depression.¹⁰⁶ Similarly, future studies should explore the use of extant assessments of motor slowing assessments in depression, eg, Salpetiere motor scale,¹⁰⁷ to provide unique insight into motor slowing in CHR populations. Additionally, future studies would benefit from continuing to validate the SMAP-R, and to expand the scale further, using the number of available motor assessments. For example, future validation studies would benefit from incorporating instrumental sensorimotor assessments (eg, actigraphy, velocity scaling, force variability, postural sway, electromyography)⁸² as these tests are sensitive to subtle perturbations and tap into mechanisms overlapping with those addressed by several of the self-report subscales/facets. In addition, it will be important to validate future gesture items with rater-based inventories (eg, Tulia)^103,108 and catatonia with instruments such as the Catatonia Rating Scale, Northoff Catatonia Scale, or Bush-Francis Catatonia Rating Scale ^31,33,34 as more work continues to clarify the role of this domain in those at CHR for psychosis. In terms of diverse clinical samples, the current study shows some important potential of a self-report motor scale to provide transdiagnostic insight into sensorimotor abnormalities and levels of physical activity. Future work is needed to further explore this possibility with large and clinically diverse samples to further examine the sensitivity and specificity of the scale. In the absence of longitudinal data, it is unknown whether this particular Sensorimotor Abnormality subscale relates to clinical course or ultimate conversion to psychosis. Finally, the scale assesses several sensorimotor signs featured on the National Institute of Mental Health (NIMH) Research Domain Criteria (RDoC) initiative’s new Sensorimotor Domain.^22,71,90 Promising results in the present study hint the SMAP-R may serve as a useful transdiagnostic tool.

Supplementary Material

Supplementary material is available at Schizophrenia Bulletin online.

Supplemental Figure 1. Factor analyses and data structure for the validation analytic dataset: A. Depicts the intercorrelation structure of the items, B. Depicts the Cattell scree plot of the intercorrelation matrix, C. Depicts the factor analyses compared to simulated data in a scree plot, D. Depicts the factor structure for both a 2 factor (Sensorimotor Abnormalities and Physical Activity Subscales) and 3 factor solution (Dyskinesia facet, Coordination Abnormalities facet, and Physical Activity Subscale).

Supplemental Figure 2. Factor analyses and data structure for the exploratory analytic dataset: A. Depicts the intercorrelation structure of the items, B. Depicts the Cattell scree plot of the intercorrelation matrix, C. Depicts the factor analyses compared to simulated data in a scree plot, D. Depicts the factor structure for both a 2 factor (Sensorimotor Abnormalities and Physical Activity Subscales) and 3 factor solution (Dyskinesia facet, Coordination abnormalities facet, and Physical Activity Subscale).

Funding

This work was supported by the National Institutes of Mental Health (MH094650, MH112545, MH103231, MH094650, VM; 5R01MH112613-03, 3R01MH112613-02S1, and 5R01MH112613-02, LME; 5R01MH112612-03, 5R01MH112612-02, and 1R01MH112612-01).