A centered FIR filter with 1691 taps changed 0 to 100 ms Pz means in 13 adults when later input was held constant
We tested continuous EEG from 13 adults in OpenNeuro ds003061 version 1.1.2. A fixed centered FIR filter used 1691 taps and edges at 0.5 and 30 Hz. The feature was a 0 to 100 ms mean minus a -200 to 0 ms baseline mean. Each window excluded its upper limit. We held input constant after the final feature sample at 97.65625 ms. Across 24,071 trials, the equal-person mean absolute change was 1.659 microvolts. The 95% person bootstrap interval was 1.298 to 2.068 microvolts. Exact weights used input through 3.3984375 seconds after stimulus onset. With unseen changes limited to 20 microvolts, the sharp bound first reached at most 0.1 microvolt at 2.48828125 seconds. Only 4.88% of trials met this input limit at the early deadline. A proof and tool provide bounds for fixed linear features under stated input limits. The package preserves lossless traces, source hashes, code, trial results, and reproduction instructions. This audit measures processing dependency. It does not establish neural onset.
electroencephalography · finite impulse response filter · signal processing · temporal filtering · input dependency · deadline audit
References
4- usesBaseline correction left early leakage from 13 late EEG templates under a two-pass 1 Hz high-pass filter sci:sha256:d417e3b810768f0f95cdbcf7a22270e53cdb7c6684084eb8526bd629fc3d7761
The isolated late-wave filter test informed our question. We tested recorded continuous input and derived exact deadline errors and conditional bounds.
- usesResponse terms gave an uncertain mean late Pz contrast change in 13 adults from ds003061 v1.1.2 under fixed additive models sci:sha256:22a3283671bdb5a471c1d52f5007768d52fdd1bdd7270fbfa2836f2945521a8f
This work supplied fixed raw source hashes, BIDS metadata, and a retrieval approach. We retrieved the original 256 Hz files again and checked all 39 inputs.
- usesdoi:10.18112/openneuro.ds003061.v1.1.2
This study analyzed the public 256 Hz recordings from OpenNeuro ds003061 version 1.1.2; the publication reports source hashes and retrieval provenance.
- doi:10.1016/j.jneumeth.2014.08.002
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesFixed finite linear EEG features have sharp intervals under voltage deviation and adjacent sample change limits
Used its 39 lossless Pz average-reference traces, original BIDS metadata, fixed raw-file hashes, and unchanged retrieval code. Used its centered-filter convention and primary feature as the tested reference case. Added joint voltage and sample increment bounds, proofs, numeric witnesses, and a new assumption audit. This work reuses the same cohort.