The 95% interval included zero for the background removal change in the CLN3 versus volunteer maximum frequency contrast (4–13 Hz; 95 ds008896 records; Fz reference; 80 retained seconds per person)

We tested whether fixed background removal changed posterior EEG maximum frequency in public CLN3 and volunteer records. We used all 32 CLN3 and 63 volunteer records from OpenNeuro ds008896 version 1.0.1. The main rule used Fz reference and the first 20 retained four-second epochs per person. We compared raw and adjusted maximum frequency in 4–13 Hz. The adjusted curve used a weighted log spectral line fitted outside this band. The CLN3 minus volunteer contrast was +0.065 Hz before removal and −0.944 Hz after removal. The paired contrast change was −1.008 Hz. Its 95% person bootstrap interval was −1.965 to +0.033 Hz, which included zero. This establishes neither a population change nor equivalence. The method moved maxima upward in 54 of 95 records. All six planned checks are reported. The 6–13 Hz rule gave similar raw and adjusted contrasts. A fixed constructed-spectrum test changed the contrast sign when only a common tilt changed. A proof establishes adjusted curve invariance to a positive power law gain with fixed fit bins, fixed weights, rank two, and exact arithmetic. The fitted line and band maximum do not identify a biological background or periodic oscillation. The public data lack ages and clinical fields. The group contrasts are unadjusted. All 95 fresh source files passed hash checks. An independent calculation checked all 5,558 retained epochs. The full rebuild and repeat gave identical input and output hashes. The package preserves spectra, metadata, source object versions, retrieval code, all results, and repeat instructions.

CLN3 disease · EEG · maximum frequency · spectral background · aperiodic spectrum · bootstrap interval