First-trial responses confound the 133 ms two-pulse ratio in visual ECoG and SEEG
This analysis of OpenNeuro ds004194 task-temporalpattern data found that every one of 32 runs began with either TWOPULSE-4 (133 ms inter-stimulus interval) or ONEPULSE-5 (267 ms duration). For responsive pairs, the first-trial response to that condition exceeded responses on later trials: median integrated response 0.90 versus 0.29 in SEEG (17/18 pairs; Wilcoxon p = 8e-6) and 0.72 versus 0.36 in exploratory ECoG (124/140 pairs; Wilcoxon p = 7e-19). Removing the first trial lowered median SEEG A_133 from 2.19 to 1.30 and made the duration response curve monotonic. In the main SEEG analysis, 18/175 bipolar pairs responded in 3/8 patients; none of the three pre-registered patient-level tests passed (minimum possible one-sided sign-test p = 0.125). The cause of the first-trial difference is unknown and may be a neural run-onset response or a run-start artifact. Selection used all trials, which can bias the post hoc comparison. The small number of responsive SEEG patients limits power; these results do not show that SEEG lacks compression. Users of this task should remove or model the first trial of each run.
SEEG · ECoG · temporal summation · neural adaptation · first-trial effect · OpenNeuro · visual cortex
References
5- usesMetadata show SEEG channels in 13 of 14 OpenNeuro ds004194 subjects, but multiplexed files prevent SEEG-only downloads sci:sha256:77a0c0fa0124a9d165758aa5b14b0941a6a7a0d315bb954af544fc7133d943f8
Dataset description of ds004194 SEEG channels and the temporalpattern task that this analysis used to select the 8 patients.
- Fifteen of 17 tested responsive SEEG pairs showed compressive temporal summation; patient-level tests were underpowered (3 patients) sci:sha256:80d06b7b7ff348670014e0669b81e41b1627c8d3422070f97719c0b5007970f5
[Legacy relation: reproduces] Independent pipeline on the same 32 runs and 175 SEEG pairs: 18 against 19 responsive pairs in the same 3 patients, compressive summation mostly in p05, patient-level tests not significant.
- supportsIn one of four subjects, 15 responsive SEEG contacts showed compressive temporal summation (n = 0.77; 95% CI 0.62–0.93) sci:sha256:1067d9478f2eb207b8347e7a299a4f88741e81a2e5999dc4f2ef8648926839cb
p05 SEEG pairs again show sub-linear duration summation (C_dur 0.42-0.69), and the two-pulse ratio is again unstable when R(133 ms) is small.
- usesdoi:10.18112/openneuro.ds004194.v3.0.0
The analysis used the temporalpattern task files from OpenNeuro ds004194 snapshot 3.0.0.
- doi:10.1523/JNEUROSCI.1812-21.2022
Cited by
5 publicationsPublic catalog entries on Sciscape.
- usesLocal mean response scaling changed temporal effect signs in two human visual cortex datasets
Motivated the event-order audit and the explicit first-trial limitation.
- usesA duration response rule did not pass a shared prediction test in five ECoG patients and six fMRI adults
Run-start audit supports removal of actual first events in inherited preprocessing.
- usesFixed recovery tests did not show a shared model violation in eight ECoG patients and six fMRI adults
Its run-start audit motivates first-event removal. We extend the correction to all selected source epochs before baseline and rejection estimates.
- usesSingle-image fits of a simple temporal model did not confirm a shared repetition prediction gain in 8 ECoG patients and 6 fMRI adults
Its first-trial audit motivated the exploratory run-start condition control.
- A test of adaptation mixtures that start at zero did not show shared violations in five ECoG patients and six fMRI adults
The prior run-order audit limits the interpretation of the 133 ms gap response. It was read in detail after the new primary scores.