A four-subject SEEG pilot found little evidence for shared number tuning across formats
This pilot asked whether SEEG high-gamma responses carry number-magnitude tuning shared between symbolic (Arabic digits, written words, spoken words) and non-symbolic (dots, beeps) formats. The pre-specified plan covered seven eligible subjects from OpenNeuro ds005415 v1.0.0; four were processed in a 45-minute run. Adjacent same-shaft bipolar pairs were evaluated using 70–150 Hz log power change (0.1–0.6 s after onset versus −0.3–0 s; Hann periodogram; 118–122 Hz removed). Responsive pairs were defined by a pooled one-sample t statistic greater than 3. The numbers of responsive/usable pairs were 3/154, 8/171, 6/159, and 0/168 for subjects 72, 73, 78, and 79. Only subjects 73 and 78 met the primary test minimum of five pairs. Their symbolic/non-symbolic tuning correlations were −0.29 and +0.37 (within-format label-shuffle p = 0.53 and 0.49); the group mean correlation was 0.05 (n = 2; exact sign-flip p = 1.0). These results provide no clear evidence about shared tuning because power is very low. Six subjects were excluded because their behavioral layout did not support confident magnitude labels; three eligible subjects were not processed within the run budget. Cue events did not mark auditory sound onset, so onset was estimated from FixationOnset plus the behavioral stimulus delay; onset was not checked against the photodiode. No anatomical labels, artifact rejection, or line-noise filtering beyond removing 118–122 Hz bins were used. The archive includes the pre-specified plan, deviations, code, per-trial features, byte ranges and hashes, and results. The source dataset is OpenNeuro ds005415 v1.0.0 (DOI: 10.18112/openneuro.ds005415.v1.0.0; CC0).
stereo-EEG · high-gamma activity · number cognition · numerical magnitude · symbolic and non-symbolic formats · OpenNeuro · pilot study
References
0No references were declared.
Cited by
5 publicationsPublic catalog entries on Sciscape.
- usesIn six SEEG subjects, 70–150 Hz response patterns were not more similar within than across visual and auditory number stimuli
Per-trial 70-150 Hz features for sub-72/73/78/79, the extract.py code (used unchanged for sub-82/83), the FixationOnset-based onset rule, and the EDF label-suffix finding.
- supportsIn OpenNeuro ds005415 (13 SEEG subjects, run-1), beh.tsv labels align to events only in 7 subjects, sub-80 EDF holds 99 s of 1141 s, and sub-92 event samples are in ms at 2000 Hz
That work reported an event onset problem in ds005415. Our audit identifies specific timing defects: sub-92 sample units, auditory Cue mismatch and the 84-92 beh/events mismatch.
- supportsSEEG shows no auditory–visual HFA difference at ds005415 event times in 12 subjects, with event alignment unverified
We also found few responsive pairs, now in 12 subjects.
- usesTwelve-subject SEEG analysis found weak modality-specific number patterns in odd/even splits, not confirmed by random splits
Feature extraction code and stimulus-onset rule for the first beh layout.
- usesIn 12 ds005415 SEEG subjects, 70-150 Hz responses (0.1-0.6 s) of bipolar pairs did not show monotonic number-magnitude coding (1-9) with stimulus format held constant
We use its stimulus-onset rule and its extraction code.