Fixed linear models showed no clear gain from 200–275 ms N2pc voltage in offline response time estimates for later blocks in 39 screened ERP CORE adults

We tested whether fixed 200–275 ms N2pc voltage improved offline response time estimates in ERP CORE. Linear models used task history, posterior baseline voltage, and eye voltage. They fitted the first four blocks and scored later blocks from the same adults. The primary sample had 39 adults, 4,008 fitting trials, and 3,865 test trials. The target was log response time. Each adult had equal weight. Adding N2pc voltage reduced mean squared test error by 0.0263%. The 95% interval from 2,000 whole-adult draws with model refitting was −0.2966% to +0.1919%. The interval included zero in all ten fixed secondary tests. The result applies to screened correct trials and this fixed linear model. It does not prove absent EEG information. Full-record two-way filtering limits the test to offline use. The SCI archive retains the result tables, model parameters, test predictions, resampling draws, figure, local code, hashes, and retrieval instructions; external source files and derived input caches are excluded because of conflicting distribution terms. The archived original package reports an offline replay matching all 71 numerical files and fresh raw downloads reproducing the first and last source records.

N2pc · event-related potentials · EEG · response time · visual search · linear models · offline prediction