The reproducible dimensionality of the human intracranial evoked response grows without detectable ceiling up to 480 trials

How many reproducible dimensions does the evoked response of a human brain contain? Perturbational complexity measures summarize the richness of a brain's response in a single number. Companion work by this author shows that what these indices track is the response's reproducible dimensionality (R-dim): the number of orthogonal response components that replicate across independent halves of the recorded trials, estimated fully out-of-sample. In simulated recurrent networks, R-dim converges with trial count to a finite ceiling that characterizes the system. This preprint asks whether the human ceiling becomes visible when far more trials are available than standard protocols provide. Using the open COGITATE intracranial-EEG dataset (visual stimulation, up to 720 trials per subject), and with the complete analysis plan, exclusion rules, two hypothesis tests and their four-cell interpretation matrix preregistered with a public timestamp before any data file was opened, R-dim was estimated across a trial grid of 30 to 480 trials in 32 usable subjects of 38 (58-194 intracranial channels; 245-698 valid trials). Main result: R-dim keeps growing to the largest trial counts measured. The sealed deceleration test was negative (high-n increments, median +0.58, not significantly smaller than low-n increments, median +0.72; p = 0.32), while top-range growth was robustly positive (p = 1.8e-5). The preregistered verdict cell reads "no detectable ceiling up to n_max". The verdict is unchanged under a 1-30 Hz analysis band (between-band Spearman rho = 0.986) and holds separately within the n_max = 240 and n_max = 480 subgroups. A four-condition audit shows the estimator does not manufacture this pattern: on response-free noise R-dim falls toward zero with increasing trials; on time-shifted real-structure data it stays below 1; on a 3-component synthetic response it saturates immediately; and on a 25-component response with decreasing amplitudes it grows steadily, recovering only about 24% of the true dimension at 480 trials. That structure-matched control reproduces the human curves closely, supporting the reading - offered explicitly as calibration-based conjecture, not measurement - that the human evoked response contains many weak reproducible components, plausibly placing its true dimensionality in the tens. A dissociation reinforces this: the level of R-dim scales with electrode coverage (Spearman rho = 0.58 with channel count), but the growth at high trial counts does not (rho = 0.03). Implication for practice: single-number complexity comparisons across studies with different trial counts are confounded by design; R-dim curves, not points, are the comparable object. Together with previously reported TMS-EEG results, trial-scaling growth of reproducible dimensionality is now documented in 62 brains across two modalities and two independent datasets, in each case under a preregistration sealed before the data were seen. Files: full preprint (HTML), figure 1 (per-subject curves and increment scatter), figure 2 (four-condition estimator audit), and a reproducible code archive containing all analysis scripts, preregistrations, per-subject numerical results, and the verification script that recomputes every number claimed in the paper. No human subject data are redistributed.