# Coverage decision before response calculation

2026-10-09. Applies the finalized amendment without changing thresholds or response definitions.

The complete extract (ZIP SHA-256 c64e8857f0062acadecea190209bfbe96de16c2897862c4654f34a93e43cc154) supplies extraction code, byte-range checksums and an explicit statement that all source spike-array entries within requested windows were retained for every unit. An included independent check reports agreement for ten deterministically selected units; these remote checks are supplied evidence, not independently rerun here.

All six SST probe LFP timestamp ranges start near 3.64–3.67 s and end near 9816.11–9816.31 s; all six wild-type ranges start near 3.70–3.72 s and end near 9937.98–9938.00 s. They enclose the requested SST windows (9212.8594–9778.89404 s envelope) and wild-type windows (9365.53193–9704.12674 s envelope). These are supporting probe acquisition bounds. Only endpoints are supplied; internal timestamp gaps and uninterrupted unit-level spike detection are not verified. No per-unit observation intervals or invalid-times table are available. Running-wheel timestamps are not used as spike acquisition bounds.

Under the finalized amendment, use coverage_status=assumed_from_acquisition_metadata, conditional on uninterrupted unit observation in these event windows and no unreported gaps. Do not use first/last spikes to include/exclude units or trials. Retain all listed isolated 10 ms events unless structural validation detects a discrepancy. SST inference remains conditional on this assumption. Wild-type remains descriptive, because each level has 15 trials. Do not interpret absent annotations as proof of no gaps, or a silent valid extracted trial as a missing observation.

Archive checksum, unit alignment, window coverage, sorted finite times and QC mapping will be checked locally before calculating responses. No response calculation preceded this decision.
