sciscape Publications
← All publications
eess.SP · Signal Processing

A six-bit buffer code cut float32 payload size by 95.1% and raised mean balanced accuracy by 0.99 points in eight fixed offline SEEG decoders

We tested a packed signal message on public SEEG data from eight people. The fixed models used earlier trials. The test used all 674 later trials and 1348 buffers from the parent study. Each pair buffer had mean removal, fourfold filtered sampling, and an adaptive six-bit code. Its message included eight float32 steps. The payload used 782 bytes, compared with 16000 bytes for eight native float32 bipolar signals. Mean balanced accuracy changed from 70.69% to 71.67%. The planned one-sided 95% upper bound on mean accuracy loss was -0.24 percentage points, below a two-point limit under independent normal person effects. The code changed 54 native decisions. All 36064 tested constant contact-offset windows kept the same message exactly. A fixed four-bit control gave 63.18% mean accuracy. Independent numerical checks and complete repeat runs passed. The package preserves exact input samples, pinned source ranges, messages, models, code, results, and figures. These reused, event-aligned trials support an offline reading method. They do not establish online control, device power savings, or a writing interface benefit.

signal compression · adaptive quantization · stereoelectroencephalography · brain-computer interface · balanced accuracy · signal transport

Admitted 0 citing publications

References

6
  1. usesZero-centered 8-bit waveform codes reduced frozen SEEG task-state accuracy from 70.7% to 50.6% in eight recordings sci:sha256:f11a99582db0d0fff28986ef433b9cde1661cb1668cb5772f0a78262582c2b8c

    Its fixed waveform-range results and training-centered follow-up informed the choice to remove the current buffer mean and send a current-buffer scale. The new code uses reduced-rate adaptive messages. No data or code from that study enters the new scores.

  2. usesSample clipping raised mean balanced accuracy from 56.5% to 63.8% under simulated 1 mV, 10 ms contact faults in eight SEEG recordings sci:sha256:76e13d01e49a34d263825687c9046fad559ed2ce66b54a45f100bcc1237bee88

    Uses the source metadata and pinned EDF range recipes through the immediate parent. A new scalar EDF reader checks one freshly retrieved trial per person against these range hashes and exact contact samples.

  3. usesEight SEEG pairs classify pre-response and rest windows at 70.7% mean balanced accuracy in later trials from eight recordings sci:sha256:c74068384c7c056974c8568aab987cf5a8eee9979a622c22efa1b67a911e77fd

    Uses the inherited candidate features, trial rules, chronological split with one gap trial per person, training-only eight-pair panels, scaler, and logistic model states. All native features and probabilities match exactly.

  4. usesFourfold filtered sampling reduced mean balanced accuracy by 0.58 points in eight offline SEEG decoders sci:sha256:f6f8bca9acbea141638cc5f4a86edba485bc8db75a281ef10abe65f1fc04225a

    This parent supplied all exact selected later-trial digital contact samples, fixed models, source metadata, range recipes, feature code, retrieval code, and known mean-removal findings. The new work tests an actual packed adaptive message and all used contacts at both constant-offset signs.

  5. usesdoi:10.18112/openneuro.ds004473.v1.0.2

    The analyzed recordings are from OpenNeuro ds004473 version 1.0.2.

  6. doi:10.1088/1741-2552%2Facae0a

Cited by

0 publications

Public catalog entries on Sciscape.

No incoming citations on Sciscape.