category00: A Pre-Registered, Hardware-Ready Re-Test of the Simplest Possible Categorical Task

We report category00: a hardware-ready re-test of the simplest possible categorical discrimination task an organoid-simulation closed-loop framework can pose — two electrodes, one driven by each of two labels, a single pulse each, with no continuous magnitude, no shortcut, and no natural axis for a decoder to exploit without genuinely learning the distinction. This exact design, in this project's prior simulation work, returned zero measurable learning signal across every tested drive level, connectivity density, and learning rate. That negative result stands, unrevised. What is new here is not the task; it is the discipline applied to re-asking it on real tissue. Before writing any code, we wrote down two competing hypotheses and a decision rule to distinguish them: either (H1) Three-Factor STDP genuinely does not drive categorical learning at this scale, in which case a real culture should replicate the simulated null; or (H2) the null result is an artifact of this project's simplified single-compartment, homogeneous-threshold organoid model, in which case real tissue — with its cell-type diversity and genuine biophysical complexity absent from the simulation — might show a signal simulation could not produce. The decision rule was fixed before any hardware run: a single favorable session does not count, only a result repeated across independent sessions distinguishes a real effect from noise. We report three pieces of hardware-readiness work, each verified against this project's real simulation engine. First, a live panel and automatic video-recording system, built before any calibration code, documenting any real-tissue session start to finish regardless of outcome. Second, a calibration script sweeping membrane resistance and reporting raw separability directly; run in simulation, it replicates the prior null exactly across every tested resistance from 10 to 80 MΩ. Third, a three-stage seed-qualification protocol, redesigned around repeatability rather than generalization to a weak signal (this task has no weak variant), which correctly halts after Stage 1 when no seed shows raw separability rather than wasting computation on a foregone conclusion. We report category00 as ready for its first real-tissue session in the same bounded sense used throughout this project's other work: the code is verified, the decision rule is fixed in advance, and the outcome — whichever hypothesis it supports — is actionable immediately. Source code, tooling, and the pre-registration document (written before any code): https://github.com/Metin1558/Category00