Two-bit rank codes raised later pulse response AUC by 0.019 over one-bit rank codes after five fit pulses in eight SEEG sessions
Two-bit sample rank codes gave mean later pulse response AUC 0.920. One-bit rank codes gave 0.900. The gain was 0.0191, with a descriptive 95 percent person bootstrap interval from 0.0103 to 0.0284. The test used OpenNeuro ds004696 v1.0.1, eight SEEG sessions, 86 stimulation sites, and five early fit pulses followed by five separate later test pulses per site. The main buffers contained 983 samples at 20 to 500 ms after each pulse and matched before-pulse samples. Two-bit sample payload was 246 bytes, compared with 3,932 bytes for float32, a reduction of 93.7 percent. Full voltage gave higher AUC, 0.939. Four rounded increasing transforms kept rank AUC fixed, but changed some near-tied sample codes. Clipping reduced two-bit AUC to 0.897. The SCI package retains native buffers, packed messages, templates, source hashes, source audit records, figures, and result checks; GPL analysis code is omitted. This is an offline method test on a reused clinical cohort. It does not measure continuous false alarms, device power, or stimulation benefit.
SEEG · sample rank codes · pulse response · AUC · brain-computer interface · signal processing · OpenNeuro ds004696
References
5- supersedesTwo-bit rank codes raised later pulse response AUC by 0.019 over one-bit rank codes after five fit pulses in eight SEEG sessions sci:sha256:9beffcb4cfc76c5de25c1270ffe91442f3029bf29f5ad1878e22923535a650e9
The first archive was made while the final file inventory was still being written. It omitted MANIFEST.json. This corrected archive includes the completed inventory. All scientific inputs, code, figures, and results are unchanged.
- usesPacked SEEG signs reduced float32 waveform bytes by 96.9 percent and mean response AUC by 0.028 in eight sessions at 20 to 500 ms sci:sha256:ff570172c3d14a876b961b937e57455653c6bcdd6eb25ec3d78df84d4ef7a144
This study uses the fixed eight-person candidate cohort, chronological five-fit/five-test pulse split, native selected bipolar waveforms, fit-only pair selection energies, baseline reader code, and hash-checked source retrieval recipe. It compares new one-to-four-bit sample rank readers with the preserved sign and voltage readers. It does not add an independent cohort.
- usesdoi:10.18112/openneuro.ds004696.v1.0.1
The analysis uses the native trial voltage and compact source-derived inputs from this exact dataset version.
- usesdoi:10.1523/JNEUROSCI.2201-22.2023
This is the source study for the limbic electrical-stimulation recordings analyzed here.
- doi:10.1103/PhysRevLett.88.174102
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesRice run coding reduced mean framed two-bit rank bytes by 60.3 percent on later pulses in eight SEEG sessions
We reused all 24 compact input files, the corrected pulse mapping, contact selection, two-bit rank reader, and source retrieval code. We reproduced its two-bit AUC from the same inputs. The new test measures framed lossless transport bytes and exact score recovery.