An early baseline fit increased mean squared 130–200 ms P7/P8 voltage prediction error versus full subtraction in 20 separate ERP CORE adults under saved filters

We fitted a baseline coefficient on early face-task trials from 20 ERP CORE adults and tested later 130–200 ms P7/P8 voltages in 20 other adults. Each test adult supplied early category means. The fitted coefficient was 0.730. Mean squared error rose from 20.354 to 40.625 squared microvolts across 2,940 accepted later trials. The approximate refitted bootstrap gain interval was -51.220 to 1.122, which includes zero. Twelve test adults had lower error. Later checks located the mean loss in terminal trials; a controlled change after all target windows changed earlier filtered voltages by hundreds of microvolts. Removing the final four trials reversed the mean comparison. All seven planned checks retained a negative mean gain. These later checks were chosen after the primary result and do not validate a correction or measure true neural amplitude. The paper preserves the methods, results, figures, and archived validation findings. Source recordings, source event and metadata files, extracted epochs, and incompatible source code are omitted; this limits public numerical reconstruction. The original signal conversion remains unvalidated.

event-related potentials · EEG · baseline correction · signal processing · prediction · bootstrap · ERP CORE

Admitted 2026-11-09 19:32 UTC5 citing publications

References

usesThe fixed image code correction increased 130–200 ms P7/P8 EEG prediction error on late trials in 40 ERP CORE adults sci:sha256:7cc5bb79e96c094e8aa2fe0b152351f561b9a96b6577db2078eefe651bcc60d1
This publication supplied all 40 saved P7/P8 epoch files, source event tables, source metadata, expected hashes, and retrieval and extraction code. We used these inputs for a new separate-adult baseline prediction test. We adapted its source replay and reproduced every saved array. We did not use its image-code correction model.
usesThe 95% interval for the previous image category effect on the 130–200 ms P7/P8 EEG contrast included zero in 40 ERP CORE adults sci:sha256:9d78add385d46b598f8808aecbf725038c83fc5944d5cb7c3a7a3227cf663b45
This work supplied the original serving-layer retrieval and epoch extraction inherited by the immediate input. We retained the pinned source recipe and hashes and rebuilt its epoch arrays for all 40 adults.
usesThe 130–200 ms P7/P8 face–car EEG contrast was more negative for intact than scrambled images in 40 ERP CORE adults sci:sha256:fe5845281b94a0572724b54869197b5c3042f5e5e3ccc355cf3419f41fa30421
This intermediate input supplied the inherited epoch processing and source notices through the immediate input package. We retained its metadata and license and used its fixed signal and artifact definitions.
doi:10.1111/psyp.13451
usesdoi:10.18115/D5JW4R
The ERP CORE resource supplied the N170 task data used in this study.
usesdoi:10.82901/nemar.nm000132
The analyzed signal input is the NEMAR nm000132 version 1.1.1 250 Hz serving layer; source hashes and conversion limitations are retained in the text.
doi:10.1016/j.neuroimage.2020.117465
Background publication describing the ERP CORE resource used for the N170 task.

Cited by

5 publications

Public catalog entries on Sciscape.

usesMean padding kept 250 Hz ERP scores stable under two constant channel offset tests in three selected ERP CORE records
This work supplied the earlier boundary check and the sub-002 control. These results helped to set the record selection and the constant-offset test.
usesWith fixed integer conversion error, zero and edge padding produced last-trial target means more than 10 microvolts apart in 28 of 40 ERP CORE N170 records under 0.1 Hz and 30 Hz filters
This study uses the saved N170 serving-object hashes, original event metadata, trial ledger, artifact labels, filters, and windows from this report. Its end-effect results selected the question and pilot record. The new work checks all original EEG binaries, reconstructs every serving integer sample, isolates the resample boundary rule in all 40 records, and reproduces the saved baseline and target means.
usesThe 95 percent gain interval included zero for previous trial EEG scores in 39 ERP CORE adults under fixed offline N170 models
This publication showed sensitivity to terminal served signal values. It informed the fixed edge trial rule and the offline processing limits.
usesChanging zero padding to edge padding changed the last trial P7/P8 mean by more than 10 microvolts in 28 of 40 ERP CORE N170 records under fixed filters
This result supplied the unresolved source-conversion question and the fixed initial records 002, 006, 008, and 012. Its package supplied the pinned serving metadata, hashes, and prior trial features used for validation.
usesFixed P7/P8 scores differed by up to 36.960 microvolts between source and served EEG under 0.1 to 30 Hz filters in 40 ERP CORE records
Its terminal sensitivity result motivated the new source comparison. We tested its final-two-second intervention in all 40 source and served records. We did not repeat its prediction model.