The association between alterations in motor and cognitive dimensions of schizophrenia-spectrum disorders: A systematic review

) share common neural underpinnings, highlighting the necessity for a thorough exploration of the connections between these areas. This relationship is crucial, as it holds potential significance in unraveling the underlying mechanisms of SSD pathophysiology, ultimately leading to advancements in clinical staging and treatment strategies. The purpose of this review was to characterize the relationship between different hyper and hypokinetic domains of motor alterations and cognition in SSD. We systematically searched the literature (PROSPERO protocol CRD42019145964) and selected 66 original scientific contributions for review, published between 1987 and 2022. A narrative synthesis of the results was conducted. Hyper and hypokinetic motor alterations showed weak to moderate negative correlations with cognitive function across different SSD stages, including before antipsychotic treatment. The literature to date shows a diverse set of methodologies and composite cognitive scores hampering a strong conclusion about which specific cognitive domains were more linked to each group of motor alterations. However, executive functions seemed the domain more consistently associated with parkinsonism with the results regarding dyskinesia being less clear. Akathisia and catatonia were scarcely discussed in the reviewed literature. The present review reinforces the intimate relationship between specific motor alterations and cognition. Identified gaps in the literature challenge the formulation of definitive conclusions. Nevertheless, a discussion of putative underlying mechanisms is included, prompting guidance for future research endeavors.


Introduction
Schizophrenia-spectrum disorders (SSD) encompass a vast array of alterations that span different dimensionsmost notably psychotic symptoms, but also cognitive and motor alterations (Kirkpatrick et al., 2014).Cognitive alterations have been extensively described in SSD, affecting a wide range of cognitive functions and being related to the extent of recovery after the disorder's onset (Halverson et al., 2019).The National Institute of Mental Health (NIMH) established the Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) consensus which identified cognitive subdimensions affected in SSD that appeared as segregated from each other and these included: working memory, attention/vigilance, verbal learning and memory, visual learning and memory, reasoning and problem solving, speed of processing, and social cognition (Green et al., 2004).
Though motor alterations were recognized more than a hundred years before contemporary psychiatry, they were mostly absent from SSD literature until recently (Walther et al., 2020;Whitty et al., 2009).Evidence has shown these alterations are part of the neurodevelopment disorder in SSD and not only side effects of D2-receptor-blocking agents (Koning et al., 2010;Pappa and Dazzan, 2009).Furthermore, specific hyper and hypokinetic motor alterations (formerly referred to as Extrapyramidal Symptoms or EPS) such as tardive dyskinesia (TD) and parkinsonism also seem to have prognostic value for clinical outcome in psychosis (van Harten et al., 2014) while others, such as dystonia, akathisia, and catatonia, have less clear significance (Pieters et al., 2022).Another construct encompassing motor alterations is the group of 'neurological soft signs' (Bachmann et al., 2014).These have also been shown to be associated with worse prognosis (Pieters et al., 2022).However, they are not commonly assessed in clinical practice and incorporate phenomena beyond motor function such as sensory integration and primitive reflexes.Waddington and Youssef (1986) were among the first to note an association between TD and cognitive impairment in a chronic patient population.This was hypothesized to be caused by some structural neural vulnerability predisposing to both alterations.The understanding of such common factor could potentially unveil the underlying mechanisms implicated in worse outcomes in psychotic disorders.Indeed, more recently, research has shown that brain structures related to motor and cognitive processing, especially the basal ganglia, are anatomically and functionally closely related (Leisman and Shafir, 2016;Obeso et al., 2014).However, the heterogeneity of methodologies to study motor and cognitive alterations, together with changes in prescribing practices of antipsychotic agents over the years, renders a complex picture from which to draw conclusions.Yet, identifying the specific motor and cognitive subdimensions within SSD, as well as understanding their interplay, can be beneficial for two main reasons.Firstly, it can facilitate the development of more clinically informed research, shedding light on the neurobiological mechanisms underlying SSD.Secondly, it can aid in the improvement of diagnosis, staging, and treatment interventions for individuals with SSD.
The primary objectives of this review in patients with SSD are: (i) To examine the associations between motor alterations (hyperkinetic, hypokinetic, and catatonic symptoms) and cognitive alterations; (ii) To identify which motor and cognitive subdimensions alterations are related and to determine the strength of those associations.
We hypothesize that significant associations with cognitive function will be present for all motor alterations subdimensions.These associations should be stronger in relation to executive functions for their role in goal-directed behavior.

Search strategy, study eligibility and selection procedure
Our protocol, based on the PRISMA guidelines (Moher et al., 2009), including its detailed search strategy, is described below and can be retrieved from https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42019145964.Searches were conducted through MED-LINE, Embase, and PsycINFO online databases and complemented through independent hand searching of reference lists of included full text articles.The searches focused on titles, abstracts, and keywords.Three sets of terms were used in combination, separated by the Boolean term AND, referring to: 1) clinical population (psychosis OR psychotic OR schizophreni* OR schizoaffective OR delusional); 2) cognition (cogniti* OR neurocogniti* OR memory OR attention OR "executive functions" OR "processing speed" OR IQ OR "intelligence quotient" OR reasoning OR learning); and 3) motor alterations (motor OR movement OR EPS OR extrapyramidal OR dyskine* OR akathisi* OR parkinsoni* OR dystonia OR catatonia OR catatoni* OR catalep*).The rationale for the chosen terms was, respectively, the following: 1) terms associated with schizophrenia-spectrum disorders; 2) terms reflecting the MATRICS consensus domains, and general intelligence; and 3) terms covering extrapyramidal and catatonic symptoms.Even though the term 'extrapyramidal symptoms', or EPS, has fallen into disuse, we will employ it as the reviewed articles might still make abundant use of it.The searches were limited to peer reviewed articles, research based on humans, papers written in English and publication date from 1980 (date of DSM-III publication, hence granting higher diagnostic reliability) to December 2022.The results were stored in EndNote.Two researchers, blinded to each other's decisions (B.M. and L.M.), performed the article selection based on the following inclusion criteria: 1) original clinical studies, hence not including reviews; 2) cognition and motor variables assessed by trained professionals, or through computerized tests in the case of cognition; 3) reporting of measures of association between motor and cognitive variables (both measures of strength and significance of association) or any test statistic and level of significance resulting from group comparisons where motor and cognitive variables are used -one to define groups and the other as the variable being compared.Duplicated papers were identified and removed.A final selection of papers was made based on discussion of divergences between the two researchers (B.M. and L.M.) with support from the other co-authors.

Data extraction
Variables extracted from the final selection of papers included: author names, title, publication year, study type, setting, recruitment dates, diagnostic categories included, number of participants, demographic variables, cognitive subdimensions assessed and specific tools used, motor subdimensions assessed and specific tools used, and statistical analysis (including level of significance).

Quality assessment
Each paper underwent a quality assessment following the NIH Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (National Institutes of Health, 2014) and obtained score was then transformed into a percentage.

Registration of review protocol and subsequent changes
The review protocol was registered in advance with the International Prospective Register of Systematic Reviews (PROSPERO), protocol CRD42019145964.
After group discussion, three changes were adopted during the implementation of the protocol and added to PROSPERO.First, articles that solely relied on instrumental measures (e.g., finger-tapping) were also excluded as the association of these methodologies with EPS is not straightforward and also because our search was not designed to cover them.Secondly, quality assessment was changed from Cochrane risk of bias tool to the NIH tool mentioned above as it seemed better fitted for the purpose.Thirdly, Paulsen et al. (1994) extensive review of literature allowed us to use their results instead of reviewing each of the previous papers.

Results
We selected 66 articles for narrative synthesis (see Fig. 1 for selection flow diagram).The selected articles are summarized in Tables 1-4, according to motor subdimensions and subgroups of interest (studies in the context of first episode psychosis and spontaneous movement alterations, i.e. non-medicated individuals).
The quality assessment of the included studies revealed modest scores on average, with half of cross-sectional studies scoring approximately 5-6 out of 10.The remaining studies were evenly split below and above this range.The main limitations identified were the lack of appropriate control for confounders and absence of blinding for the assessors.Among the selected studies, only seven were considered to have a longitudinal design, and they achieved a slightly better score of 63 % on a 14-point scale.The main limitation for these studies was a loss to follow-up rate exceeding 20 % (see Supplementary Table 1).Sample sizes ranged from 10 to 1310 participants, with an average of 169 (TD and mixed motor alterations studies had the higher average sample sizes -194 and 197, respectively; parkinsonism and FEP/drug-naïve studies had the lowest average sample sizes -108 and 75, respectively).Main findings are summarized in Fig. 2.

Tardive dyskinesia
The review article by Paulsen et al. (1994), which included 21 of the originally selected studies, supported a correlation between TD and cognition.We found 30 subsequent articles focusing on TD that confirmed the same trend (see Table 1), including five studies conducted on a geriatric population (Berry et al., 2007;Byne et al., 1998;Karson et al., 1990;Karson et al., 1993;Quinn et al., 2001).
The instrument used to assess TD was, in most cases, the AIMS, with TD defined according to the Schooler & Kane criteria (Schooler and Kane, 1982).The studies were more heterogeneous regarding how TD was categorizedwhether as a continuous or categorical variableleading to different statistical approaches.The two studies with higher sample sizes were derived from the same clinical trial sample and applied an analysis of covariance (considering TD dichotomously).
The included neuropsychological subdimensions varied across studies.Nevertheless, most research tried to grasp cognitive function in general with the use of tests or batteries covering the main cognitive subdimensions, often relying on the Mini-Mental State Examination (MMSE).A smaller group of studies focused on executive functions and found associations with TD, namely attention, task switching and setshifting abilities (Spohn and Coyne, 1993;Waddington et al., 1995;Waddington et al., 1993).The larger sample-size studies also indicated a stronger relationship between TD and executive functions (attention and immediate memory) than other neurocognitive domains (Hui et al., 2017;Liang et al., 2022).Another study that specifically assessed executive functions in elderly patients showed this subdimension to be a strong predictor of TD, further adding a topographic nuance to the finding as the result was not applicable to limb-truncal TD (Quinn et al., 2001).Overall, only three studies did not find any association between neuropsychological variables and TD (Barnes et al., 1995;Miller et al., 2005;Pourcher et al., 1993).
Two studies using a longitudinal design were retrieved.A 10-year       cohort study (Waddington and Youssef, 1996) found that cognitive impairment and TD seemed to develop together within the same time interval and not predict each other (at least within a 5 and 10-year window).Another study, with a 6-month follow-up found that TD was associated with less improvement in cognition over time (Caroff et al., 2011).

Parkinsonism
While some studies focused specifically on parkinsonism (see Table 2), this motor subdimension was often assessed together with other movement disorders (see Table 3; first episode psychosis and spontaneous movement-disorders studies are addressed separately).Six studies reported no significant association between parkinsonism and any of the cognitive subdimensions assessed (Ahmed, 2013;Byne et al., 2000;Chen et al., 2001;Docx et al., 2012;Silver and Shlomo, 2001;Sullivan et al., 1994).Studies with positive findings reported correlations between parkinsonism and different cognitive outcomes which varied according to the study's methodology (i.e.general cognition score, verbal learning, visuo-spatial abilities, spatial working memory and frontal lobe functions).The only longitudinal study in chronic SSD (Hwang et al., 2012) showed that parkinsonian symptoms were crosssectionally associated to IQ and were predicted on a 6-month followup by a baseline response inhibition task (r = 0.72, p < 0.001).

Akathisia, catatonia and mixed motor alterations studies
Only 4 studies reported clinician rated scores of akathisia in relation to cognitive performance.The study with the largest sample did not report any association between akathisia and cognition (Ahmed, 2013).Hwang et al. (2012) found a cross-sectional link between akathisia and IQ but no associations between baseline cognition and follow-up akathisia.The study by Sachdev et al. (1996) identified an association between akathisia and a symbol-substitution task.Kim et al. (2002) found that individuals with akathisia (or higher akathisia scores) had worse mental control (a very sensitive measure of attentional dysfunction).
We only selected three studies that addressed catatonia in chronic medicated patients (Bark et al., 2005;Chen et al., 2001;Docx et al., 2012).The only significant association was found in the study by Bark et al. (2005) which examined set-shifting ability differences between a group with 8 patients diagnosed with catatonic schizophrenia and a group of patients with paranoid schizophrenia.
Kim et al. conducted a series of studies analyzing objective EPS in relation to subjective cognitive dysfunction.In three separate articles they showed that akathisia and parkinsonism are related to the overall score and specific sub-dimensions of a subjective cognitive-perceptual dysfunction scale.Zhornitsky et al. (2011) also showed an association between subjective cognitive capacities and subjective motor alterations.
A group of studies, predominantly more recent, investigated different motor subdimensions together, sometimes employing a single combined scale for the assessment of extrapyramidal symptoms such as the EPS rating scale or the Drug-induced EPS scale (DIEPSS).These studies on mixed motor alterations more frequently reported a lack of association between motor and cognitive variables (Chen et al., 2001;Silver and Shlomo, 2001;Sullivan et al., 1994).

First episode psychosis and spontaneous movement alterations studies
Seven articles presented studies on the first episode psychosis (FEP) population, dealing with a range of different movement disturbances.Cuesta et al. (2014) and Pareek et al. (2010) investigated spontaneous parkinsonism in FEP in relation to a mixed group of cognitive subdimensions in the former and executive functions in the latter.Cuesta et al., using a longitudinal design, found a predictive value of baseline spontaneous parkinsonism over cognitive dysfunction on a 6-month follow up, but found no cross-sectional differences at baseline between patients with and without spontaneous parkinsonism.A notable 21-year  follow-up study (Peralta et al., 2024) found no predictive value of spontaneous motor alterations over cognition, but conversely found premorbid IQ to predict follow-up motor alterations.With a medicated FEP sample, Lindgren et al. (2022) also found baseline parkinsonism to predict 1 year global, verbal and visuomotor cognition.Pareek et al., using a cross-sectional design, exclusively investigated executive functions through the Wisconsin Card Sorting Test and found a significant positive correlation between spontaneous parkinsonism and perseverative errors.Compton et al. (2015) and Kindler et al. (2019) investigated dyskinesia in FEP.Neither study found a significant association between dyskinesia and cognition (a neurocognitive battery was used in the former study and pre-morbid IQ assessment in the latter).As a side note, the study by Kindler et al. enrolled only 10 patients with FEP, but included 45 subjects with Clinical High Risk (CHR) for psychosis.In that subgroup, a trend was found for the negative association between dyskinesia and premorbid IQ.The studies by Colomer et al. (2017), Cuesta et al. (2018), andCompton et al. (2015) investigated the associations between catatonic symptoms and cognition in FEP patients.Only the study by Colomer et al. with a drug-naïve samplefound catatonia to be negatively correlated with cognition (working memory and visual learning), with statistical significance.Cuesta et al., with a similar approach, were not able to find an association.Conversely, these authors found significant associations between EPS and cognitionparkinsonism with visual memory, and akathisia with visual memory  and speed of processing.Two other studies investigated spontaneous movement disorders in patients with longer durations of psychotic disorders (Fenton et al., 1994;Molina et al., 2016).These studies investigated treatment naïvepatients and found worse general cognition in patients with spontaneous dyskinesia (orofacial) and parkinsonism, respectively.

Discussion
In this review, we analyzed 66 articles that investigated clinical assessments of hyper and hypokinetic motor alterations, including catatonia, in relation to neuropsychological variables within the context of SSD.Most of the reviewed studies reported significant associations between the included motor subdimensions and cognition.Most reported correlations fall within the weak or moderate range.Cognition is still frequently approached as a single entity, limiting the ability to conduct a more in-depth analysis of the specific cognitive subdimensions that correlate with motor alterations.The contribution of antipsychotics or other medications as a potential confounding factor to the reported correlations is still not completely addressed, but evidence from well controlled studies and medication-naïve samples indicates an effect that extends beyond confounding.
Overall, our findings support cognition and the array of motor alterations in SSD as fundamental components that span from phenomenological to neurobiological approaches to these disorders.This perspective aligns with the systems neuroscience theory of disconnectivity in corticobasal circuits in SSD, particularly in relation to the parallel loops connecting the basal ganglia, cerebral cortex, and cerebellum, which are implicated in affective, cognitive, and motor functions (Bernard et al., 2017;Obeso et al., 2014).This formulation might, however, be too imprecise to be informative for effective translation into clinical impact.In fact, the array of motor alterations encompasses a range of phenomena which, despite sharing certain neural substrates, are likely characterized by distinct interactions with cognitive function.We will delve into these nuanced interactions in the subsequent discussion.

Tardive dyskinesia
The association between TD and cognition gathered significant attention as a research topic in the 1980s and early 1990s.The review article published in 1994 by Paulsen et al. (1994) shed light on the methodological limitations of the previously published articles, including poorly defined samples, inadequate neuropsychological assessments, inadequate statistical testing, poor control for confounders, and excessive use of cross-sectional designs.The analysis of subsequent studies conducted in the present review revealed that although there has been an improvement in methodological quality over the years, there is still a lack of high-quality studies dedicated to addressing this question, and many of the limitations identified in the 1994 review still apply.Significant issues related to the samples used in the studies need to be addressed, which compromises the generalizability of the results to the SSD population.A considerable number of studies primarily relied on acute inpatients or even chronically institutionalized patients, potentially selecting a population with a worse prognosis.Furthermore, the choice of cognition subdimensions and the instruments used to assess them often involved simplistic assessments (e.g.MMSE), which might have limited the sensitivity to capture cognitive variability associated with TD.
Research into the pathophysiology of dyskinesia in SSD helps understanding its association with cognitive alterations.Emerging evidence underscores thalamo-cortical hyperconnectivity as a pivotal factor for this motor alteration, exhibiting a notable correlation with aggravated cognitive deficits, particularly pertaining to attention and processing speed (Chen et al., 2019;Walther et al., 2017).Structural neuroimaging investigations have revealed a noteworthy diminution in the volume of both the putamen and the cerebellum in association with tardive dyskinesia (Sakreida et al., 2022), both of which have long been regarded as quintessential motor-regulating structures, yet their involvement in cognitive functions is increasingly apparent.
We observed a trend in more recent studies, which consistently demonstrated negative results regarding the association between TD and cognition, particularly when other motor alterations were also assessed.Interestingly, these studies tended to report a higher frequency of associations between cognition and parkinsonism rather than dyskinesia.This shift could stem from earlier research that predominantly focused on TD in older patients, often treated with typical antipsychotics.Age and antipsychotic dosage are acknowledged confounders when examining motor-cognitive associations.The distinction between typical and atypical antipsychotics' effects on this link remains unclear.Yet, it is plausible that increased use of atypical antipsychotics, coupled with shorter treatment durations, reduced TD severity while maintaining observable parkinsonism in more recent samples.Additionally, potential publication bias may underlie studies concentrating solely on TD, favoring statistically significant results for publication.

Topographic distinctions
Considering the somatotopic organization of the basal ganglia (Nambu, 2011), it is important to explore the potential variations in the associations between TD and cognitive function based on the topography of the dyskinesiaorofacial TD (OTD) vs. limb-truncal TD.The finding that OTD is more associated with cognition may be biased due to the higher occurrence of OTD compared to limb-truncal TD in the reviewed literature.Notably, two high-quality studies from the reviewed pool indicate that both OTD and limb-truncal TD may be associated with cognitive alterations but through different subdimensions (Quinn et al., 2001;Wu et al., 2013).In the future, a more precise understanding of how the basal ganglia contribute to cognitive functions may help elucidate the distinct associations with altered movement in different body parts.

Parkinsonism
Most studies assessing parkinsonism have found a weak to moderate association with worse cognitive outcomes in different subdomains.Across studies, deficits in executive functions, learning, and visuospatial abilities were associated with higher parkinsonism levels.Interestingly, one large-sample negative study did not include the assessment of executive functions (Ahmed, 2013).Consistent with these findings, the literature on Parkinson's disease suggests a rationale for an association between parkinsonism and frontal lobe functions (Rodriguez-Oroz et al., 2009).Despite their scarcity, studies on parkinsonism within SSD also reveal the implication of structures with active roles in cognitive processes involved in goal-directed behavior.Molina et al. (2018) not only described an association between parkinsonism and a set of cognitive skills but also investigated how parkinsonism related to structural connectivity alterations.Their results pointed to prefrontal cortex-caudate dysconnectivity (higher fractional anisotropy) being related to parkinsonism score.This association held true even in patients with limited exposure to antipsychotic medication.A recent imaging study also pointed to an association between parkinsonism and volume reduction of the left caudate nucleus and the motor cortex in medicated patients with schizophrenia, a pattern that differed from dyskinesia, as noted above (Sakreida et al., 2022).Hence, at least in a subgroup of patients with SSD, alterations to the caudate nucleus development might give rise to both parkinsonian symptoms and difficulties in planning goal-directed behavior.It is important to note, however, that the association between executive functions and motor output such as parkinsonism (but potentially other motor categories) might in part be caused by an overlap of these constructs (motor planning).These new insights into the distinct associations of parkinsonism and dyskinesia with specific brain structures in SSD might help understand their correlations with particular cognitive functions.However, studies designed to investigate this question further are still warranted.
The evidence from the few longitudinal studies retrieved, combining medicated and non-medicated samples, suggests that under antipsychotic medication, parkinsonism primarily correlates cross-sectionally with cognition.However, the same motor alteration in drug-naïve samples might hold prognostic value by predicting cognition upon follow-up (Cuesta et al., 2014).Nevertheless, conflicting evidence exists.Interestingly, in a 21-year follow-up, spontaneous parkinsonism did not predict cognition, but baseline IQ predicted follow-up parkinsonism (Peralta et al., 2024).This highlights the complexity of the correlations at hand, in close proximity with other mediating and moderating variables over time.
The role of motor speed bias in the association between cognition and parkinsonism does not appear to be consistently supported.Two separate studies that controlled for this factor yielded opposite results (Fervaha et al., 2015;Palmer et al., 1999).Moreover, since certain neurocognitive assessments rely on motor speed, further research is needed to evaluate potential effect modification and/or bias.This factor may also be relevant in regard to dyskinesia (Eberhard et al., 2006).Additionally, assessments of parkinsonism may be influenced by other clinical factors such as negative symptoms or other causes of psychomotor retardation.Properly controlling for these in the clinical setting can prove challenging as the constructs themselves are ill-defined (Walther and Strik, 2012).Consequently, the use of instrumental and objective assessments for motor outcomes could be particularly advantageous in this field (van Harten et al., 2017).
There was a lack of studies addressing social cognition.Parkinsonism seemed the motor alteration potentially with stronger negative correlations with that cognitive domain, butas recent network analysis studies suggestwith the mediation of neurocognition (Monteleone et al., 2021;Moura et al., 2021;Silver and Shlomo, 2001).Despite the accumulating evidence regarding the interconnectedness of social cognition and motor function during human development (Kenny et al., 2016), this area remains understudied in the field of psychosis.

Akathisia
The mechanisms underlying this hyperkinetic syndrome are still poorly understood (Musco et al., 2020).Simultaneously, the subjective discomfort that is part of the presentation of akathisia makes it more challenging to interpret cognitive dysfunction in this context.The reviewed articles exhibited significant variations in cognitive assessments, which limited the comprehensive interpretation of results.However, discrepancies were evident (e.g.attentional deficits were found in one-third of the studies assessing attention).Overall, further clarification is needed on this topic.

Catatonia
According to Bark et al. (2005), the association between catatonic symptoms and impairment in set-shifting abilities could be attributed to dysfunction in the ventral prefrontal cortex.Consistent with this finding, a study assessing stereotypies (not included in this systematic review due to the device-assisted assessment) found a positive correlation between perseverative errors on the WCST and the severity of stereotypies (Morrens et al., 2006).Studies specifically designed to investigate this question are warranted as the cited evidence is based on exploratory research.Dean et al. (2020) conducted a study on patients with SSD and a previous history of catatonia.They included a broader range of cognitive assessments and hypothesized that cognitive functions relying more on motor performance, such as verbal fluency and processing speed, would be significantly more impaired in these patients compared to those without a history of catatonia, which was the case.
The different approaches mentioned above highlight the heterogeneity of conceptualizations of catatonia, with an impact in research questions (i.e.catatonia as a disorder of will vs. a motor disorder; Walther et al., 2019).

First episode psychosis and spontaneous movement alterations studies
Studies conducted during the early stages of psychosis show the presence of interrelated motor (parkinsonism, akathisia, catatonic symptoms) and cognitive symptoms (executive functions, memory, and attention), in medicated and non-medicated individuals.These symptoms may reflect early pathophysiological processes involving corticalstriatal circuits with aberrant dopaminergic transmission (Howes et al., 2009).
Overall, parkinsonism appears to be the motor syndrome most consistently associated with cognition in the early stages of psychotic disorders.However, the number of studies retrieved was very limited, with a modest total number of participants.
Spontaneous dyskinesia and parkinsonism in antipsychotic-naïve samples are variables of interest for further studies.Our findings suggest an association between motor and cognitive alterations in this subgroup of patients, which aligns with a previous review on spontaneous movement disorders in first episode psychoses (Pappa and Dazzan, 2009).Spontaneous motor alterations have mixed evidence of predicting cognition on follow-up studies.However, it is challenging to find unmedicated individuals with established SSD, so the field is gaining traction in the at-risk population (not reviewed in the present article).For example, Mittal et al. (2010) reported a correlation between dyskinetic movements and cognitive alterations (verbal comprehension, perceptual organization, and auditory memory), with these symptoms predicting at-risk individuals transitioning to psychotic disorders on follow-up.In the present review, our findings suggest that the relationship between hyper and hypokinetic motor changes and cognitive alterations is independent of medication use.Nonetheless, medication might influence this connection in certain vulnerable individuals later on.Exploring these factors in early psychosis could shed light on distinguishing primary pathological processes from medication-related effects, offering valuable insights for diagnosis and prognosis.

Subjective reports of cognitive and motor dysfunction
A weak relationship emerged between subjective motor (parkinsonism and akathisia) and cognitive alterations indicating that also in the context of insight motor and cognitive (specific) alterations might be correlated.Very few studies were however retrieved on this topic.Subjective motor alterations seemed strongly related to objective motor alterations.The opposite apparently happens between subjective and objective cognition (Medalia and Thysen, 2008).Of note, there have been recent developments in the study of subjective experiences of catatonia (Brandt et al., 2024), which could, perhaps in the future, also integrate subjective and objective appraisals of cognitive function.

Limitations
This study is limited by the lack of a meta-analytic approach.Given the considerable heterogeneity among the studies, the use of metaanalytic statistics would have been challenging to apply.Therefore, we opted for a narrative synthesis following a systematic search.
The exclusion of certain research fields also limits the scope of the present review.We did not include other types of motor symptoms and assessments (e.g., neurological soft signs, oculomotor studies) nor studies that focused on the 'at-risk-mental state' for psychosis.This choice was made to maintain a concise analysis.Additionally, the term 'sensorimotor' (whose relevance stems from the use within the RDoC initiative) was not a search term, which could have excluded articles of interest from this review.
An important limitation, intrinsic to the topics at hand, concerns the construct overlap between different motor alterations.Thus, the precise delineation of motor alterations should be approached with caution, as these motor syndromes may not be entirely distinct from one another due to conceptual or neurobiological overlaps (Peralta et al., 2010).Another limitation from the included studies is the insufficient explanation regarding the persistence of motor alterations termed as 'tardive.'Although 'tardive' is commonly associated with dyskinesia, it is not consistently applied to describe other motor syndromes which may not always resolve following the discontinuation of the causative medication (Aquino and Lang, 2014).Moreover, in most samples the response of motor alterations to the reduction of antipsychotic medication was not assessed.This lack of information could obscure the differentiation between tardive and non-tardive motor alterations, particularly concerning their cognitive consequences.

Conclusions and recommendations for future studies
The reviewed studies point to significant connections between specific motor subdimensions and cognition in SSD, regardless of medication.Further in-depth research is necessary to elucidate detailed information on the specific cognitive and motor subdomains that are associated.Investigating these associations longitudinally would provide insight into their prognostic value and variations over the course of the disorder alongside with (dis)continuous medication use.Furthermore, we emphasize the following points as suggestions for enhancing future studies: (i) incorporating assessments that encompass clinician ratings, instrumental evaluations, and subjective accounts of motor alterations, this way contributing to more comprehensive and valid assessments of the motor domain; (ii) formulating hypotheses that guide the selection of cognitive assessments.Drawing from the current review, it seems important to include a comprehensive set of assessments evaluating executive functions, as well as processing speed and motor speed.Addressing the presented gaps in the literature and suggestions to overcome them could yield a more profound understanding of the interactions between specific motor and cognitive functions within SSD.

Declaration of competing interest
None.

Table 1
Studies concerning tardive dyskinesia as the motor alteration of interest.

Table 1
(continued on next page)

Table 1
(continued ) Mini Mental State Examination; NART -National Adult Reading Test; OTD -Orofacial Tardive Dyskinesia; RBANS -Repeatable Battery for the Assessment of Neuropsychological State; sczschizophrenia; TD -Tardive Dyskinesia; TMT -Trail Making Test; WAIS -Weschler Adult Intelligence Scale; WCST -Wisconsin Card Sorting Test; WMS -Weschler Memory Scale.Studies concerning parkinsonism as the motor alteration of interest.

Table 3
Studies concerning akathisia, catatonia or a mixed set of motor symptoms.

Table 4
Studies concerning first episode psychosis or drug-naïve samples (spontaneous movement disorder).