Removing 114 to 126 Hz from 70 to 150 Hz power changed mean SEEG state accuracy by −0.36 percentage points on later trials in eight recordings with fixed pairs and refitted readers
We tested whether removal of 114 to 126 Hz changes a SEEG task-state reader on later trials. We used OpenNeuro ds004473 version 1.0.2, eight recordings, eight fixed bipolar pairs per person, and 674 later test trials. The complete 70 to 150 Hz power feature gave 72.13 percent mean balanced accuracy. Removal of 114 to 126 Hz and a new model fit gave 71.77 percent. The paired mean change was minus 0.36 percentage points. The descriptive 95 percent patient bootstrap interval was minus 1.14 to plus 0.53 points. This interval does not establish equal accuracy. Keeping the complete-band model fixed gave a larger mean change of minus 1.60 points. Equal-width omissions near 100 and 140 Hz also had smaller mean effects after a new fit. The package preserves exact digital inputs, source version pins, features, models, all test predictions, three figures, code, and independent checks. All 2,014 retained trials were rebuilt from pinned source ranges. Every saved input field matched exactly. The study used reused clinical trials and event-defined states. It did not test signal origin, live control, transfer between days, or stimulation.
stereoelectroencephalography · SEEG · spectral power · frequency band removal · task-state classification · logistic regression · brain-computer interface · reproducibility
References
6- usesFiltering before fourfold sample removal gave 70.6% mean balanced accuracy in later trials from eight SEEG recordings sci:sha256:8bc11c5e9496aba0a89fbf8535edb5ce1336d5a00cc9df2b9795ddfb7b8ddde4
This direct archive supplied exact 500-sample digital input buffers, fixed selection files, source metadata, pinned byte range records, and MIT source retrieval code. We used its source-rate inputs and added fixed spectral omission tests.
- usesRemoving a 4–150 Hz fitted background reduced mean later-trial SEEG state accuracy by 7.1 percentage points in eight patients with eight fixed pairs each sci:sha256:c3469f7de479fc287d2128fe8ba0836892261cc3bf0dc5c837eb721c6ef303b1
This intermediate archive supplied compact native digital inputs through the direct archive. Its source records and citation chain are preserved. Its fitted spectral background conditions are not used.
- usesEight SEEG pairs classify pre-response and rest windows at 70.7% mean balanced accuracy in later trials from eight recordings sci:sha256:c74068384c7c056974c8568aab987cf5a8eee9979a622c22efa1b67a911e77fd
This study defines the inherited trials, eight early-selected pairs, rest and before-response windows, and native 70 to 110 Hz reference. We keep its selection and later-trial split. The new tests are conditional on those pairs.
- usesA stationary Gaussian variance correction with 5 percent shrinkage reduced SEEG decoding accuracy for 20 percent simulated sample loss in eight recordings sci:sha256:495cd9ff22cc042e1f138d100dbf95110efd76545b22997f0bb2f49092bdf3ae
This archive is part of the verified digital source chain. Its native buffer and source range records reached this study through the direct archive. We preserve its source metadata record for provenance.
- usesdoi:10.18112/openneuro.ds004473.v1.0.2
The analyzed digital buffers and source metadata are pinned to OpenNeuro ds004473 version 1.0.2; its recorded license is CC0.
- doi:10.1088/1741-2552%2Facae0a
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesA frequency derivative projection left a 0.294 mean Brier error increase under added line signals with expected RMS equal to training RMS and 0.5 Hz fundamental shifts in eight SEEG records
We used the exact public digital voltage buffers, source metadata, and fixed selection records from this archive. Its omitted-bin method is the secondary comparison.
- usesAdding spectral slope increased mean later Brier error in eight SEEG recordings with fixed pairs and penalty 1
This archive supplied the exact digital buffers, fixed selection files, source metadata, and retrieval code. The new study computes slope and ratio features and fits new readers on those inputs.
- usesA degree 5 projection reduced added Brier error by 0.294 compared with degree 1 in eight offline SEEG readers under constant ±0.5 Hz fundamental offsets and expected added RMS equal to training RMS
We used its inherited 114 to 126 Hz power omission comparator and exact short-buffer benchmark through the direct parent package.