Recovery is the measurement that matters.

The question everyone asks at the moment of a head injury is whether it is a concussion. It is the wrong question to build a company around. Not because it does not matter, but because it is the crowded, contested part of the field, and because the more consequential question comes days and weeks later, when someone has to decide whether the athlete or the soldier is ready to go back. That decision is made, today, almost entirely on judgment. When it is made too early, the downside is not a missed game. It is second-impact syndrome, a rare but frequently catastrophic event that follows a return to contact before the brain has recovered.^[1] The measurement that would make that decision less of a guess does not really exist yet. That is the one we are trying to build.

The sideline is the crowded part

Acute detection at the sideline already has capable tools. The SCAT5, King-Devick, VOMS, and Sway Medical all live at the moment of injury, and Sway is already FDA-cleared and phone-based. Competing there means fighting for a few points of accuracy on terrain that is already well occupied. It also means competing on a problem whose ground truth is genuinely thin. Concussion has no objective gold standard. There is no blood test or scan that settles it, so every detection tool is validated against another imperfect human judgment. In independent review, even computerized neuropsychological testing has shown sensitivity well below 100 percent.^[2] A tool can only be as good as the label it is measured against, and at the moment of injury, that label is soft.

Recovery is the cleaner problem

Recovery monitoring inverts almost every one of those disadvantages. The decisions are high-stakes and, right now, subjective, which means the unmet need is real rather than incremental. The usage is repeated rather than one-and-done, because recovery is tracked over a series of sessions, not captured in a single check. And the ground truth is cleaner, because we can anchor to something concrete: an independently made clinical clearance date. A clearance date is a decision a qualified clinician actually recorded, on a real calendar, for a real person. It is a far more honest thing to validate against than a contested diagnosis at the chaotic moment of a hit. Detection can stay the headline. Recovery is the part worth building a company on.

The right comparison is you, last week

The other reason recovery is the better frame is that it changes the question from one we cannot answer well to one we can. Asking whether a person is normal runs straight into the enormous natural variation between healthy people. Asking whether a person has returned to their own normal sidesteps that entirely. Every reading is measured against their own prior sessions, so the comparison controls for who they are and isolates what has changed. This is why the system captures a baseline first and treats every later session as a movement relative to it.

We measure that movement across four independent signal families, captured on a standard phone or tablet camera rather than dedicated lab equipment. The oculomotor channel tracks how the eyes move, which commodity cameras can now do at sub-degree accuracy.^[3] The autonomic channel reads pulse from subtle color change in the face, a technique with a long validation history,^[4] and the pupillary light response, which phone-based pupillometry has been shown to capture.^[5] The postural channel measures sway through on-device body-pose tracking,^[6] building on work showing that mobile devices can quantify postural control in a clinically meaningful way.^[7] Reaction time is the fourth. No single one of these is the answer. Fusing them, and reading each against a personal baseline, is the bet.

The dataset is the instrument

A claim worth making about recovery requires a particular kind of evidence: a longitudinal, multimodal, outcome-labeled cohort that runs from a pre-injury baseline, through the acute window, into recovery, and ends at a confirmed clinical clearance. That dataset is what would let anyone show that a recovery score tracks a real return to health. It is also the one thing in this field that cannot be reconstructed after the fact. You cannot go back and capture a baseline for an injury that already happened. So the discipline we hold ourselves to is to record the clearance date and the outcome from the very first session, not as administrative overhead, but as the measurement itself. The instrument is not only the software in the room. It is the labeled history it accumulates.

What we are not claiming

This essay describes a direction and a bet, not a finished result. The validation that would turn the recovery score into something a clinician should rely on is ahead of us, not behind us, and we would rather say that plainly than imply otherwise. Today Chronic Trace is a wellness and research tool. It is not a medical device, and it is not cleared to diagnose or to make a clearance decision for anyone. The work is to earn the harder claims with data, in that order. Choosing to measure recovery rather than detection is the first part of doing that honestly, because it points the company at the question whose answer we can actually prove, and at the dataset that would prove it.