Re: Clarifying comparative claims in digital cognitive assessment: Assessing concurrent validity in a shared cohort

In response to correspondence from Cognivue, Inc. regarding our publication (Jannati et al., 2025), we clarify that our analysis employed a rigorous paired-sample design using the same 930 participants from the Bio-Hermes-001 study for both the Digital Clock and Recall® (DCR) and Cognivue Clarity® assessments. The assertion that our findings rely on “nonequivalent cohorts” or “outdated regulatory data” is factually incorrect; performance metrics were calculated de novo using the harmonized study dataset. Furthermore, we address the omission of the Cognivue Amyloid Risk Measure (CARM), noting that comparing a cross-validated predictive model (DCR) with a post hoc metric derived from the same dataset (CARM) would introduce significant bias due to overfitting. We conclude by highlighting how head-to-head comparisons in diverse cohorts align with the recently published Global CEO Initiative on Alzheimer’s Disease (CEOi) recommendations for digital cognitive assessment standards.