Category	Variables Extracted
Study Identification and Design	First author and publication year Country and geographic location Study design (prospective cohort, retrospective cohort) Single-site versus multi-site study Funding source and conflicts of interest
Participant Characteristics	Total sample size (UHR participants) Number of converters and non-converters Conversion rate (percentage) Age: mean, standard deviation, and range Sex distribution (percentage male/female) Ethnicity and racial composition UHR criteria used (CAARMS, SIPS/SOPS, Basel, other) UHR subtype distribution (APS, BLIPS, GRD) Baseline symptom severity (SIPS, PANSS, CAARMS scores) Baseline functioning (GAF, SOFAS scores) Antipsychotic medication status at baseline (%, dosage) Cannabis use (% users, frequency, lifetime exposure) Other substance use Psychiatric comorbidities (depression, anxiety, PTSD) Family history of psychosis Educational level
Neuroimaging Assessment Details	Scanner manufacturer and model Magnetic field strength (1.5T, 3T, other) Imaging modality (sMRI, fMRI, DTI, MRS, multimodal) Structural MRI sequence parameters (TR, TE, voxel size, slice thickness) Functional MRI paradigm (task-based: specific task; resting-state: duration) DTI sequence parameters (number of directions, b-values) MRS voxel location, size, and metabolites measured Brain regions assessed (specific regions or whole-brain) Analysis approach (ROI, VBM, surface-based, tract-based, connectivity) Software used for image processing and analysis Quality control procedures Head motion assessment and correction methods Artifact detection and exclusion criteria
Follow-up and Outcome Assessment	Follow-up duration (months or years) Conversion assessment method and criteria (DSM-IV/5, ICD-10/11) Assessment instruments used (SCID, CAARMS, SIPS) Assessor training and inter-rater reliability Blinding of assessors to neuroimaging results Follow-up assessment schedule and frequency Time to conversion for converters Specific psychotic disorder diagnoses (schizophrenia, schizoaffective, brief psychotic, other) Loss to follow-up (number and percentage) Reasons for attrition
Results and Effect Sizes	Neuroimaging measures: means and standard deviations for converters and non-converters Statistical test results (t-tests, F-tests, Z-scores, p-values) Effect sizes: Cohen’s d, Hedges’ g, partial eta-squared, other Confidence intervals for effect sizes Classification accuracy metrics: sensitivity, specificity Positive predictive value (PPV) and negative predictive value (NPV) Area under receiver operating characteristic curve (AUC-ROC) Likelihood ratios (positive and negative) Odds ratios and hazard ratios for conversion Multivariable model results (beta coefficients, p-values) Machine learning classification performance (accuracy, precision, recall, F1-score) Cross-validation results Feature importance or weights from machine learning models Subgroup analyses (by sex, age, UHR subtype, medication status)
Statistical Methods and Confounders	Statistical tests used Multiple comparison correction methods (FWE, FDR, Bonferroni, other) Covariates included in analyses: - Age correction - Sex as covariate - Total intracranial volume correction - Education level adjustment - Baseline symptom severity adjustment - Medication exposure correction - Cannabis use adjustment - Head motion parameters (for fMRI/DTI) Handling of missing data Power calculations and sample size justification Statistical software used
	Note: All extracted data were entered into a standardized spreadsheet and double-checked for accuracy. Any discrepancies between reviewers were discussed and resolved by re-examining the original source publication. Study authors were contacted when data were missing or unclear, with two reminder emails sent at two-week intervals if no response was received to the initial contact.