Cohort Lineage and Independence

A new paper is not automatically an independent replication. Reused cohorts, nested autopsy subsets, and consortium analyses must be tracked before judging consistency.

COH001 | Boston v2.0 multicentre tissue cohort

Recruitment: Symptomatic referral patients with MRI and tissue

Known reuse: Derivation, temporal, geographic, and full-autopsy analyses within one publication

Bias relevance: Tissue verification and referral spectrum can inflate transportability

Independence note: Validation subsets are separated but arise from the same coordinated study program

COH002 | ROSMAP

Recruitment: Religious Orders Study and Rush Memory and Aging Project

Known reuse: Boston transportability, ARTS development, arteriolosclerosis, cognition, and mixed-pathology studies

Bias relevance: Volunteer, highly educated, older cohort; repeated reuse is not independent replication

Independence note: Publications may use overlapping participants and differently filtered subsets

COH003 | Mayo nonhemorrhagic MRI-autopsy cohort

Recruitment: Community/memory presentations without hemorrhage

Known reuse: Boston v2.0 nonhemorrhagic validation

Bias relevance: Small selected pathology subset and nonhemorrhagic spectrum

Independence note: Independent setting from Boston derivation but modest sample

COH004 | Mild Stroke Study 3

Recruitment: Edinburgh/Lothian minor ischemic stroke services

Known reuse: Cross-sectional and longitudinal CVR/SVD analyses

Bias relevance: Stroke enrichment limits generalization to community aging and specific vessel pathology

Independence note: Analyses using the same cohort should not be counted as independent replication

COH005 | MarkVCID multisite consortium

Recruitment: Older adults across participating US sites

Known reuse: PSMD, free water, CVR, clinical and fluid protocols

Bias relevance: Heterogeneous sites improve transportability but outcomes often remain associative

Independence note: Sites may be analyzed independently, but consortium-wide papers share infrastructure and some participants

COH006 | ADNI

Recruitment: Research volunteers across cognitive stages

Known reuse: ARTS translation and outcome association

Bias relevance: AD-focused selection and exclusions limit arteriolosclerosis generalization

Independence note: May be external to ROSMAP development but not representative clinical care

COH007 | Edinburgh lobar ICH pathology cohort

Recruitment: Adults with spontaneous lobar ICH

Known reuse: Edinburgh CT and genetic criteria

Bias relevance: Applies to lobar ICH, not cognitive or asymptomatic presentations

Independence note: Distinct context from MRI-based Boston criteria

COH008 | Specialist CAA-ri comparative cohort

Recruitment: CAA-ri and non-inflammatory CAA specialty care

Known reuse: CAA-ri spectrum characterization

Bias relevance: Referral enrichment and incorporation of imaging into diagnosis

Independence note: Useful comparative phenotype evidence; not population accuracy

COH009 | International postmortem CAA protocol network

Recruitment: Eleven neuropathology groups

Known reuse: Consensus protocol development, appraisal, and validation

Bias relevance: Consensus improves comparability but does not eliminate sampling error

Independence note: Method paper rather than participant-outcome cohort

COH010 | Hereditary CAA family cohorts

Recruitment: Mutation-defined families, including Dutch-type CAA

Known reuse: Natural history, risk factors, and mechanistic staging

Bias relevance: Founder effects, mutation specificity, younger age, and limited numbers

Independence note: Different mutations should not be pooled without biological justification