A Computer-Assisted Tool for Objective Assessment of Extrapyramidal Symptoms

Development of an Objective Quantitative Assessment and Computer-Assisted Diagnostic Tool for Extrapyramidal Symptoms

This prospective, single-center observational study aims to develop and validate a computer-assisted tool for the objective quantitative assessment of extrapyramidal symptoms (EPS).

The study will enroll 180 participants, including 90 healthy controls and 90 participants with clinically confirmed antipsychotic-induced EPS. The healthy control group will provide reference data on normal movement, while the EPS disease group will provide data on abnormal movements associated with drug-induced parkinsonism, akathisia, or tardive dyskinesia.

Each participant will complete one study visit lasting approximately 60 minutes. During the visit, a trained psychiatrist will assess EPS using the Simpson-Angus Scale (SAS), Barnes Akathisia Rating Scale (BARS), and Abnormal Involuntary Movement Scale (AIMS). Participants will also complete standardized movement tasks while a non-contact computer vision system records and analyzes movement features. The device-based assessment will be repeated after a short rest to evaluate the stability of the measurements.

Data collected earlier in the study will be used to develop and optimize the assessment tool. After the algorithm is finalized and locked, data from subsequently enrolled participants will be used for independent validation. The study will evaluate how closely the tool's results agree with clinician assessments, as well as its ability to identify EPS, produce stable repeated measurements, and operate feasibly and safely in a clinical setting.

This is an observational study. The study does not assign treatment, change participants' medications, or use the investigational tool to make treatment decisions.

Study Overview

• Status: Not yet recruiting
• Conditions: Extrapyramidal Symptoms; Tardive Dyskinesia (TD); Akathisia, Drug-Induced; Drug-induced Parkinsonism
• Intervention / Treatment: Diagnostic test: Computer-Assisted Objective Assessment of Extrapyramidal Symptoms

Detailed Description

This is a prospective, single-center, two-cohort observational study designed to develop and validate an objective quantitative assessment and computer-assisted diagnostic tool for extrapyramidal symptoms (EPS). The study focuses on EPS associated with antipsychotic treatment, including drug-induced parkinsonism, akathisia, and tardive dyskinesia.The observational study model is classified as Other because this is a two-cohort diagnostic accuracy and tool development and validation study with a single study visit, rather than a conventional longitudinal cohort study.

A total of 180 evaluable participants are planned for enrollment. The study includes two cohorts: 90 healthy controls and 90 participants with clinically confirmed antipsychotic-induced EPS. The healthy control cohort will provide reference data for normal physiological movement and will help establish a baseline for distinguishing normal movement variability from abnormal movement features. The EPS disease cohort will provide positive reference data representing clinically confirmed abnormal movement patterns and different levels of symptom severity. The two cohorts are used as reference populations for tool development and validation and are not treatment comparison groups.

Each participant will complete a single study visit lasting approximately 60 minutes. After written informed consent, demographic and relevant clinical information will be collected. For participants in the EPS disease cohort, information on psychiatric history, antipsychotic treatment, time of the most recent medication dose, and use of medications for EPS will also be recorded.

A trained senior psychiatrist will then conduct standardized clinical EPS assessments using the Simpson-Angus Scale (SAS), Barnes Akathisia Rating Scale (BARS), and Abnormal Involuntary Movement Scale (AIMS). The clinical assessment includes observation of posture and spontaneous movement, examination of orofacial movements, standardized upper-limb and hand movements, standing and walking tasks, and assessment of subjective symptoms when applicable. Clinician ratings will be recorded separately from the device assessment, and device operators will not have access to the clinician ratings during data collection.

Participants will subsequently complete a standardized device-based assessment using a non-contact image-based computer vision system. The assessment tasks are designed as a digital representation of movement features examined during routine clinical EPS assessment. Tasks include standardized periods of sitting, upper-limb positioning and movement, orofacial movements such as mouth opening and tongue protrusion, hand and finger movements, standing, walking, and turning. The system extracts facial, hand, and body keypoint information and derives quantitative movement and kinematic features.

After the first device-based assessment, participants will have a short rest of approximately 5 minutes. When physically able and willing to continue, they will complete a second assessment using the same device, instructions, and task sequence. The repeated assessment will be used to evaluate the stability and repeatability of the device-based measurements.

The study uses a two-stage development and validation design. Approximately the first 70% of evaluable participants, including participants from both cohorts, will contribute data to the tool development stage. Device-derived movement features will be aligned with clinician-rated SAS, BARS, and AIMS results to support model training, parameter optimization, and development of objective severity grading. After the predefined development process is completed, the algorithm will be locked and no further model modification will be made for the validation analysis.

Approximately the final 30% of evaluable participants enrolled after algorithm lock will form an independent validation set. The locked system will be evaluated on these previously unseen data. The validation stage will assess agreement between device-based and clinician-rated EPS severity grading, the ability of the tool to distinguish participants with clinically confirmed EPS from healthy controls, the stability of repeated measurements, and the feasibility and safety of the standardized assessment procedure.

This is a non-interventional observational study. Participants are not assigned to treatments or medications by the study protocol. The study does not require discontinuation, dose reduction, dose escalation, or substitution of antipsychotic or anti-EPS medications. Outputs from the investigational system will be used for research purposes and will not be used to make immediate clinical treatment decisions.

• Study Type: Observational
• Enrollment (Estimated): 180

Contacts and Locations

Study Locations

• China
  • Shanghai, China
    • Shanghai Mental Health Center

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: Yes

• Sampling Method: Non-Probability Sample

Study Population

The study population will consist of a Control Group and an EPS Disease Group recruited at Shanghai Mental Health Center. Potential participants in the EPS Disease Group will be identified and screened by research physicians in relevant outpatient clinics, inpatient wards, and other institutionally approved recruitment settings. Participants in the Control Group may be recruited through institutionally approved notices or by study researchers. Eligibility will be confirmed according to the predefined inclusion and exclusion criteria. A total of 180 evaluable participants are planned, including 90 participants in the Control Group and 90 participants in the EPS Disease Group.

Description

Inclusion Criteria

A. Control Group

1. Aged 18 to 80 years, with no restriction on sex.
2. No current or past history of a psychiatric disorder meeting ICD-10 or DSM-5 criteria.
3. No current or past history of central nervous system disease, such as primary Parkinson's disease or epilepsy, and no severe physical illness.
4. No use within the 3 months before screening of medications that may cause extrapyramidal symptoms, such as antipsychotics or certain antiemetic medications.
5. Sufficient visual, auditory, and cognitive ability to understand and cooperate with the standardized device-side assessment task sequence.
6. Full understanding of the purpose and exploratory research nature of the study, voluntary participation, and provision of written informed consent.

B. EPS Disease Group

1. Aged 18 to 80 years, with no restriction on sex.
2. Diagnosis of schizophrenia or a related psychotic disorder according to ICD-10 criteria.
3. Regular treatment with one or more antipsychotic medications for at least 4 weeks.

4. Clinically confirmed antipsychotic-induced extrapyramidal symptoms meeting at least one of the following criteria:

   1. Drug-induced parkinsonism: a total score of 2 or higher on Items 1, 2, 3, 4, 5, and 9 of the Simpson-Angus Scale (SAS).
   2. Akathisia: a score of 2 or higher on the Global Clinical Assessment item of the Barnes Akathisia Rating Scale (BARS).
   3. Tardive dyskinesia: a score of 3 or higher on any one of Items 1 through 7 of the Abnormal Involuntary Movement Scale (AIMS), or a score of 2 or higher on any two of these items.
5. Sufficient visual, auditory, and cognitive ability to understand and cooperate with the standardized device-side assessment task sequence.
6. Full understanding of the purpose and exploratory research nature of the study, voluntary participation, and provision of written informed consent by the participant or, where applicable, the participant's legal guardian.

Exclusion Criteria

The following exclusion criteria apply to both groups:

1. Primary neurological or organic movement disorders that may affect the assessment results, such as primary Parkinson's disease, Huntington's disease, Wilson's disease, or sequelae of stroke.
2. Severe musculoskeletal disease, recent fracture, limb disability, or other physical conditions that prevent completion of the required assessment tasks.
3. Acute exacerbation of psychiatric symptoms with severe agitation, impulsivity, stupor, or severe cognitive impairment that prevents safe and adequate cooperation with the assessment.
4. Severe deformity of the face or tested limbs, large dark-colored tattoos, or severe non-removable occlusion that may substantially interfere with computer-vision-based assessment.
5. Poorly fitting removable dentures or frequent compensatory oral movements that may interfere with objective assessment of orofacial movements.
6. Any condition that, in the investigator's judgment, may affect participant safety, substantially compromise data quality, or make the individual unsuitable for participation.

Study Plan

How is the study designed?

Number of groups / cohorts

2

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
Control Group; Participants without current or past psychiatric disorders and without extrapyramidal symptoms associated with antipsychotic treatment. Participants will undergo standardized clinical scale assessments and two repeated device-based movement assessments during a single study visit. Data from this group will provide reference information on normal movement characteristics for development and validation of the objective quantitative assessment and computer-assisted diagnostic tool.	Diagnostic test: Computer-Assisted Objective Assessment of Extrapyramidal Symptoms; Participants will complete a standardized, non-contact, image-based computer-assisted assessment of extrapyramidal symptoms. The assessment consists of predefined sitting, orofacial, hand and finger, upper-limb, standing, walking, and turning tasks. The system processes image data to extract facial, hand, and body keypoints and derive quantitative movement and kinematic features. Each participant will complete the same assessment twice during a single study visit, separated by approximately 5 minutes, to evaluate test-retest reliability. The assessment is conducted for research purposes only. It does not assign or modify treatment, and its outputs will not be used to make immediate clinical treatment decisions.
EPS Disease Group; Participants with schizophrenia or related psychotic disorders who have received regular antipsychotic treatment and have clinically confirmed antipsychotic-induced extrapyramidal symptoms. Eligible symptoms include drug-induced parkinsonism, akathisia, and/or tardive dyskinesia according to predefined clinical criteria. Participants will undergo standardized clinical scale assessments and two repeated device-based movement assessments during a single study visit. Data from this group will be used for development and validation of the objective quantitative assessment and computer-assisted diagnostic tool.	Diagnostic test: Computer-Assisted Objective Assessment of Extrapyramidal Symptoms; Participants will complete a standardized, non-contact, image-based computer-assisted assessment of extrapyramidal symptoms. The assessment consists of predefined sitting, orofacial, hand and finger, upper-limb, standing, walking, and turning tasks. The system processes image data to extract facial, hand, and body keypoints and derive quantitative movement and kinematic features. Each participant will complete the same assessment twice during a single study visit, separated by approximately 5 minutes, to evaluate test-retest reliability. The assessment is conducted for research purposes only. It does not assign or modify treatment, and its outputs will not be used to make immediate clinical treatment decisions.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Maximum Weighted Kappa Coefficient Across Three EPS Subtype-Specific Device-Clinical Severity Classification Comparisons	The primary outcome is a single aggregated agreement metric defined as the maximum weighted kappa coefficient among three prespecified subtype-specific comparisons. Weighted kappa coefficients with 95% confidence intervals will be calculated separately for: (1) device-side drug-induced parkinsonism severity classification versus Simpson-Angus Scale (SAS)-based classification; (2) device-side akathisia severity classification versus Barnes Akathisia Rating Scale (BARS)-based classification; and (3) device-side tardive dyskinesia severity classification versus Abnormal Involuntary Movement Scale (AIMS)-based classification. The maximum of the three weighted kappa coefficients will be reported as the primary outcome in the independent test set after algorithm locking. The prespecified primary success criterion will be met if at least one of the three subtype-specific weighted kappa coefficients reaches or exceeds the predefined agreement threshold.	Single study visit, Day 1

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Test-Retest Reliability of Device-Side Repeated Acquisition Results	Participants will complete two identical standardized device-side motion acquisitions during the same study visit. The intraclass correlation coefficient (ICC) will be calculated for continuous motion features generated by the two acquisitions. Weighted kappa will be calculated for predicted classification results to evaluate short-term stability.	Two assessments during the single study visit, separated by approximately 5 minutes, Day 1
Sensitivity of the Device-Side System for Identifying EPS	Using the prespecified SAS-, BARS-, and AIMS-based clinical criteria as the reference standard, sensitivity will be calculated as true positives divided by true positives plus false negatives, multiplied by 100, with a 95% confidence interval. Analysis will use the independent test set after algorithm locking.	Single study visit, Day 1
Specificity of the Device-Side System for Identifying EPS	Using the prespecified SAS-, BARS-, and AIMS-based clinical criteria as the reference standard, specificity will be calculated as true negatives divided by true negatives plus false positives, multiplied by 100, with a 95% confidence interval. Analysis will use the independent test set after algorithm locking.	Single study visit, Day 1
Accuracy of the Device-Side System for Identifying EPS	Using the prespecified clinical reference standard, accuracy will be calculated as true positives plus true negatives divided by the total number of evaluated participants, multiplied by 100, with a 95% confidence interval. Analysis will use the independent test set after algorithm locking.	Single study visit, Day 1
Positive Predictive Value of the Device-Side System for Identifying EPS	Positive predictive value will be calculated as true positives divided by true positives plus false positives, multiplied by 100, with a 95% confidence interval. Analysis will use the independent test set after algorithm locking.	Single study visit, Day 1
Negative Predictive Value of the Device-Side System for Identifying EPS	Negative predictive value will be calculated as true negatives divided by true negatives plus false negatives, multiplied by 100, with a 95% confidence interval. Analysis will use the independent test set after algorithm locking.	Single study visit, Day 1
Area Under the Receiver Operating Characteristic Curve for the Device-Side System in Identifying EPS	A receiver operating characteristic curve will be generated using a continuous output produced by the locked device-side system and the binary clinical reference classification. The area under the curve will be calculated with a 95% confidence interval in the independent test set after algorithm locking. Higher values indicate better discrimination.	Single study visit, Day 1
Number of Participants With One or More Adverse Events During Study Procedures	The number of participants experiencing at least one adverse event during the study visit will be recorded. Potential events include falls, gait instability, muscle fatigue or soreness, anxiety, fear, agitation, refusal to undergo testing, and transient fluctuation of psychiatric symptoms. Severity, management, outcome, and relationship to study procedures will be summarized descriptively.	Throughout the single study visit, approximately 60 minutes, Day 1

Collaborators and Investigators

• Sponsor: Shanghai Mental Health Center

Study record dates

Study Major Dates

• Study Start (Estimated): July 1, 2026
• Primary Completion (Estimated): April 1, 2027
• Study Completion (Estimated): April 1, 2027

Study Registration Dates

• First Submitted: July 8, 2026
• First Submitted That Met QC Criteria: July 20, 2026
• First Posted (Actual): July 21, 2026

Study Record Updates

• Last Update Posted (Actual): July 21, 2026
• Last Update Submitted That Met QC Criteria: July 20, 2026
• Last Verified: July 1, 2026

More Information

Terms related to this study

• Keywords: Computer Vision; Computer-Assisted Diagnosis; Antipsychotic-Induced Extrapyramidal Symptoms; Objective Quantitative Assessment
• Additional Relevant MeSH Terms: Neurologic Manifestations; Central Nervous System Diseases; Nervous System Diseases; Chemically-Induced Disorders; Movement Disorders; Drug-Related Side Effects and Adverse Reactions; Poisoning; Dyskinesias; Dyskinesia, Drug-Induced; Neurotoxicity Syndromes; Pathological Conditions, Signs and Symptoms; Signs and Symptoms; Psychomotor Agitation; Tardive Dyskinesia; Akathisia, Drug-Induced
• Other Study ID Numbers: YShen-EPS

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO
• IPD Plan Description: Individual participant data will not be shared with other researchers because the study involves potentially identifiable facial and body movement data from psychiatric participants. Original images and key frame images will not be publicly shared or transferred outside the study institution. No external individual-level data sharing mechanism is currently planned.

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No