Picture a site coordinator at a CNS-focused academic medical center pulling up her CTMS dashboard on a Monday morning. She has three active neuroscience studies running, a data query backlog from last week’s remote monitoring visit, and a new protocol amendment sitting in her IRB queue. Now her sponsor’s clinical trial manager calls: the incoming AI platform will be processing continuous EEG output from enrolled subjects in real time — and she needs to confirm her site’s data infrastructure can receive and tag it according to CDISC standards. She pulls up her site’s EDC user guide. There is no field for it.
That gap — between what AI-driven platforms now promise and what site infrastructure can actually execute — is the real story behind Kordata Dynamics’ emergence from stealth, announced this month. The company is positioning itself to redefine clinical trial execution by processing neural data while aligning with the FDA’s April 2026 initiative on real-time clinical trial data. The pitch is compelling. The operational questions it creates for sites are more urgent than anyone in the press release is saying.
The Data Standardization Gap Nobody Is Solving
Kordata’s model sits at the intersection of two problems that have resisted clean solutions for years: real-time data ingestion in trial operations, and the absence of standardized pipelines for neural biomarker data. The FDA requires CDISC standards for regulatory submissions, but — and this is the part that matters operationally — no comprehensive FDA guidance exists specifically for EEG or neuroimaging data standardization across clinical trials. A site coordinator transferring continuous neural signal data into an EDC is operating without a regulatory map.
That absence has consequences that compound across a trial’s lifecycle. When source data exists in proprietary neural acquisition formats that do not map cleanly to CDISC’s SDTM domains, the deviation risk starts at data entry and grows with every downstream transfer. TMF Reference Model elements for data management plans assume structured, queryable clinical data — not high-frequency waveform output from a 64-channel EEG cap. Any site that agrees to run a Kordata-adjacent protocol without first reconciling those format questions with their data management SOP is setting up a finding at the next monitoring visit, and a harder conversation at audit.
The sponsor-side picture is equally fraught. Pharmaceutical and biopharmaceutical companies represented 64.9% of total CTMS demand in 2024, with North America holding over 45% of global CTMS revenue. Those are systems with significant incumbent infrastructure — Medidata, Veeva, Oracle — that were not architected to natively ingest the kind of continuous neural biomarker streams Kordata is describing. Medidata introduced AI-powered site monitoring tools in Q1 2024, but AI-assisted oversight of standard clinical data is a different engineering problem than real-time neural signal processing. The integration layer between Kordata’s model and existing sponsor CTMS platforms will determine whether this concept survives contact with actual trial execution.
Which raises the question site teams will be asked to answer on a practical timeline: who owns that integration, and who bears the startup cost?
Real-Time Trials Require Real-Time Readiness
The FDA’s April 2026 real-time clinical trial data initiative — which includes proof-of-concept trials — represents a genuine regulatory tailwind for platforms like Kordata’s. But regulatory appetite and site readiness are two separate variables, and the gap between them is where first-patient-in dates go to die.
Sites I work with across our network are already stretched on technology onboarding. A single protocol with a new ePRO platform, a central lab portal, and an IRT system can add 15 to 20 hours of coordinator training time before the SIV even occurs. A real-time neural data platform with its own proprietary processing requirements is not an add-on — it is a parallel operational workflow that requires dedicated staff time, hardware validation, and data transfer testing before the first subject is consented. For a site running two or three competing CNS protocols with a coordinator who has been in the role for less than 18 months, that is not a minor lift.
The NIH has recognized exactly this problem. A current NIH funding opportunity specifically targets the rigorous development and validation of digital health technology-derived biomarkers and clinical outcome assessments, acknowledging that DHT-derived endpoints are not yet operationally mature across the site landscape. Kordata is launching into a market where the regulatory aspiration exists, the scientific rationale is solid — and the site-level execution infrastructure is still catching up.
The FDA has moved on AI-assisted trial tools before. The agency qualified its first AI drug development tool for use in MASH clinical trials, establishing that AI-generated endpoints can carry regulatory weight. That precedent matters — it means Kordata’s outputs could, in principle, support label-level evidence. But the MASH qualification came with defined parameters and a clear data pathway. Neural signal processing in a real-time trial context involves continuous data streams, edge-case artifact rejection, and signal-to-noise decisions that happen faster than any CRA monitoring queue can catch. The quality management plan for a protocol running this model needs to specify who makes those decisions, on what timeline, and with what escalation path — before the first subject wears the headset.
What Changes Monday Morning
For sites being approached about neuroscience protocols that reference real-time AI monitoring or neural biomarker processing: add a technology readiness checklist to your feasibility response before you sign the site agreement. That checklist should cover EDC compatibility with high-frequency data formats, staff training hours for the AI platform, data transfer validation timelines, and an explicit line in the budget for coordinator time to manage neural data quality queries. Screen failure rates in CNS trials already run higher than sponsors project — adding a novel data collection modality without budget for the learning curve makes that problem worse, not better.
For sponsors building protocols around platforms like Kordata’s, the operational ask is straightforward: do not treat CTMS integration as a post-SIV problem. The data standardization question — specifically, how neural output maps to CDISC SDTM domains for your regulatory submission — needs to be resolved at the protocol design stage, documented in the data management plan, and tested at pilot sites before broad rollout. In our network, across studies that introduced novel data collection technology without pre-startup integration validation, the median delay to first clean data transfer has been measured in weeks, not days. That is recoverable on a Phase 2 timeline. On a Phase 3 with an adaptive interim analysis baked in, it is a critical path event.
The FDA’s real-time trial initiative and Kordata’s launch are moving in the same direction at the same moment. The sites that will execute this well are the ones that treat the infrastructure question — not the AI question — as the thing that needs solving first.
References
- FierceBiotech — “Kordata Dynamics launches with neural-focused AI clinical trial model”
- ProRelix Research — “Clinical Data Standardization in Clinical Trials: FDA Compliance in Clinical Data Management”
- Market Research Future — “Clinical Trial Management System Market Report, 2024”
- NIH — Notice of Funding Opportunity PAR-25-170: Digital Health Technology-Derived Biomarkers and Clinical Outcome Assessments
- FDA — “FDA Qualifies First AI Drug Development Tool for Use in MASH Clinical Trials”
