Shapes of direct cortical responses vs. short-range axono-cortical evoked potentials: The effects of direct electrical stimulation applied to the human brain

Clotilde Turpin (CAMIN); Emmanuel Mandonnet; F\'elix Schlosser-Perrin (CAMIN); Fran\c{c}ois Bonnetblanc (CAMIN); Hugues Duffau; Olivier Rossel (CAMIN); Riki Matsumoto; Sam Ng

REVIEW 4 major objections 5 minor 4 references

Shapes of direct cortical responses vs. short-range axono-cortical evoked potentials: The effects of direct electrical stimulation applied to the human brain

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read ACEP and DCR differ by a ~2 ms N1 delay and opposite late activity.

desk verdict Useful clinical discriminator between DCR and ACEP, but the headline delay is a one-tailed zero-crossing effect with p=0.0556 and a noisy velocity estimate; the late-component shape difference is the more defensible finding. read the letter →

arxiv 2411.16360 v1 pith:M4BFE3SR submitted 2024-11-25 eess.SP

Clotilde Turpin (CAMIN) , Olivier Rossel (CAMIN) , Félix Schlosser-Perrin (CAMIN) , Sam Ng , Riki Matsumoto , Emmanuel Mandonnet , Hugues Duffau , François Bonnetblanc (CAMIN) This is my paper · ORCID

classification eess.SP

keywords DirectElectricalStimulationEvokedPotentialsElectrocorticographyCorticalResponseAxono-CorticalPotentialElectrogenesisAwakeBrainSurgeryIntra-OperativeNeuralMonitoring

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Direct electrical stimulation of the brain can elicit two kinds of evoked potentials: direct cortical responses (DCR), recorded at the same gyrus where the cortex is stimulated, and axono-cortical evoked potentials (ACEP), recorded after stimulating white matter fibers. The paper claims these two responses can be told apart by a short delay in the onset of the N1 component (about 1.8 to 2.4 ms, with ACEP later) and by the shape of the waveform 40 to 100 ms after the stimulus artifact, where ACEP turns positive and DCR remains negative. It interprets the early delay as axonal conduction time from the white-matter stimulation site and the late difference as extra intra-cortical activity recruited only by direct cortical stimulation. If correct, these markers give a practical electrophysiological signature for judging, in real time during tumor surgery, whether the stimulation is engaging cortex directly or through subcortical fibers, which matters for interpreting mapping results and for modeling the effects of DES.

What carries the argument

The key machinery is the set of waveform metrics extracted from averaged electrocorticographic traces, chiefly the time of the first zero-crossing t_zc1 (the transition between the early positive P0 and the negative N1 component) and the signed area under the curve from 40 to 100 ms after the stimulation artifact. t_zc1 serves as a read-out of the moment the cortical population response begins, so its delay between ACEP and DCR provides an estimate of axonal conduction velocity; the late signed area serves as a read-out of whether a secondary cortical activity follows the main N1 deflection, which the paper associates with intra-cortical axon recruitment in DCR.

What would settle it

Measure ACEP and DCR from the same recording electrode while systematically moving the white-matter stimulation site closer to and farther from the cortex, and check whether the t_zc1 delay grows with distance by the expected amount. If the delay stays roughly constant, or if a short-distance ACEP shows no delay at all, the axonal-conduction interpretation is refuted.

Watch

Extended reading notes

Core claim

The paper establishes that, even though DCR and ACEP share the same canonical P0–N1 waveform, they differ in two measurable respects: the first zero-crossing t_zc1, which marks the onset of N1, occurs significantly later for ACEP than for DCR, and the signed area under the curve between 40 and 100 ms after the stimulus artifact is positive for ACEP but negative for DCR. The observed t_zc1 delay averages about 1.8 ms (median 2.4 ms), and the N1 component is wider for DCR, whose relaxation after the N1 peak is monotonic, while ACEP shows an after-positivity. The investigators interpret the early delay as axonal conduction time from the deeper white-matter stimulation site to the cortical recording electrode, estimating a mean conduction velocity of about 5.9 m/s (median 8.38 m/s), and they interpret the late-wave divergence as evidence that direct cortical stimulation recruits additional small, slow intra-cortical axons that are not engaged when the stimulus reaches the cortex through white-matter pathways.

Load-bearing premise

The conduction-delay interpretation assumes the straight-line distance between the DES site and the ECoG electrode is a valid measure of the actual axonal path length; if the true fiber path is longer or the stimulation directly excites cortex from below, the estimated velocity and the delay explanation lose their basis.

Editorial extensions

If this is right

Within a single patient, an N1 onset delay near 2 ms can classify an evoked response as an ACEP rather than a DCR, turning a simple latency measurement into a connectivity check during resection.
The mean conduction velocity of about 5.9 m/s (median 8.38 m/s) is consistent with the longer latency reported for the arcuate fasciculus, so the same delay logic can be extended to monitor white-matter pathways at larger distances.
Because DCR's N1 is wider and followed by negative activity while ACEP shows a positive after-positivity, comparing these shape features can indicate whether a recording electrode is receiving input from subcortical fibers or is on a directly stimulated gyrus.
The results imply that computational models of direct electrical stimulation should treat DCR and ACEP as distinct electrogenic processes, with an explicit axonal conduction delay for white-matter stimulation and an additional intra-cortical recruitment term for cortical stimulation.
Resolving the ~2 ms delay requires high sampling rates and no hardware filter, so lower-sampling clinical acquisition systems may systematically miss the ACEP-DCR distinction.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A natural extension is to test whether the t_zc1 delay grows linearly with stimulation-to-recording distance; if it does not, the axonal-conduction reading would need to be revised toward a local-network or synaptic-delay explanation.
Because the paper's velocity estimate uses straight-line distance rather than tract length, the true axonal conduction speed is likely lower than 5.9 m/s; tractography-based path length could correct this estimate.
The higher 50 Hz gamma power for DCR during the N1 period suggests a testable prediction that direct cortical stimulation produces stronger high-frequency cortical activation than white-matter stimulation, which could be probed with EEG or fMRI measurements.
If the ACEP after-positivity is inhibitory in origin, as the authors provisionally suggest, it could serve as a marker for the spatial extent of inhibition around a white-matter tract during awake surgery, testable with paired-pulse or pharmacological manipulations.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Quick take: this is a direct extension of the group's own DCR/ACEP framework, and the most defensible new result is not the headline delay but the late-shape difference. The delay claim leans on a single zero-crossing metric that is only significant if you accept a one-tailed test, and the conduction-velocity estimate is basically noise. Still, the paper deserves refereeing; the clinical discriminator is plausible and the recording methodology is unusually careful.

What's actually new: they systematically compare DCR and ACEP in the same patients (9 patients, 55 DCR vs 29 ACEP after amplitude threshold) and report t_zc1, N1 width, area 40-100 ms, and 50 Hz gamma. The late components are consistently different: ACEP shows an after-positivity and shorter N1, DCR stays negative and has a wider N1. The p-values for those are reasonable (0.002-0.04), and the shared-site comparisons in patients 1/5/7 support the direction. Credit also for 19.2 kHz sampling, no hardware filter, polarity alternation, and explicit artifact handling—that is the right way to record early components.

Soft spots. The headline t_zc1 result is reported as p=0.0556 in Table 1 and then called significant in the text via a one-tailed test that isn't presented as such. That's not fatal—the hypothesis was directional and a one-tailed p~0.028 is defensible—but the paper should say so plainly. More importantly, t_zc1 is a zero-crossing between P0 and N1, not an onset latency. It shifts with P0 width and N1 rise time, so it is not a clean measure of axonal conduction delay. The paper never reports P0 peak/onset latencies, and the velocity estimate (5.9 ± 43.6 m/s) shows how unstable the distance-delay relation is. The 3D Euclidean distance is a weak proxy for the actual fiber path, and some ACEPs may be direct cortical stimulation from below, as the authors admit. Also, only significant metrics are reported; several time windows look post hoc. That tempers the gamma finding too (only 50 Hz, 20-40 ms). No data or code are provided, which is common for clinical intracranial recordings but still limits independent checking.

Bottom line: the late-component distinction is real enough to be clinically useful, but the conduction-delay interpretation needs better latency measures or a larger sample. For a reader working on intraoperative ECoG/DES, this is worth a careful look. I'd send it to peer review with a request for P0 latencies, one-tailed p-values stated as such, and the full metric set.

Referee Report

4 major / 5 minor

Summary. The manuscript compares direct cortical responses (DCR) and axono-cortical evoked potentials (ACEP) recorded with ECoG in nine patients during awake brain surgery. It reports that ACEP exhibits a delayed first zero-crossing of the early P0/N1 complex (about 1.8–2.4 ms), a shorter N1 duration, a positive after-positivity between 40 and 100 ms versus a negative one for DCR, and higher gamma power for DCR at 20–40 ms. The authors interpret these differences as evidence for axonal conduction delay in ACEP and additional intracortical activation in DCR, and they argue that DCR and ACEP can be distinguished by a delay at the onset of early components and by the shape of later components.

Significance. If the reported distinctions replicate, they offer a practical way to infer the effects of direct electrical stimulation during brain tumor surgery, potentially improving intraoperative mapping. The study draws on direct human recordings at 19.2 kHz, within-patient paired comparisons, and shared recording sites for a subset of DCR and ACEP pairs, which are commendable strengths. However, the early-delay claim rests on a borderline zero-crossing metric rather than an onset latency, and the conduction-velocity estimate is extremely noisy; the late-component findings are also weakened by selective reporting. The paper explicitly acknowledges several of these limitations, which is to its credit, but the central claims need stronger statistical and interpretational support.

major comments (4)

[Table 1 and §3.2] The primary delay metric t_zc1 is called 'significant' based on a one-tailed t-test, yet Table 1 reports only p=0.0556 without labeling it two-tailed or providing the one-tailed p-value. Because the central claim of a delay in ACEP hinges on this metric, the authors should report the one-tailed p-value, state the directional hypothesis as pre-specified, and provide P0 peak or onset latencies to demonstrate that the difference is a true latency shift rather than a morphology-induced change in the zero-crossing.
[§3.2] The conduction-velocity estimate (mean 5.9 ± 43.6 m/s) is not meaningful as reported: the standard deviation is an order of magnitude larger than the mean, and the individual t_zc1 differences range from -1.71 to +5.15 ms. The paper acknowledges imprecise distance estimation, but the axonal-delay interpretation requires a positive correlation between distance and delay. Without such an analysis, or P0-based latencies, this estimate cannot support the conduction-delay explanation.
[§3.4 and Table 1] The claim that DCR relaxation is 'monotonous and positive' is contradicted by the data: Min(dDCR/dt)[50:80]ms has a mean of -0.61 ± 1.42 with p=0.8819 for the test against >0, indicating no evidence of a positive monotonic slope. This inconsistency undermines the interpretation of the Area[40:100]ms difference and should be corrected or re-analyzed.
[§2.4 and §3.1] The manuscript states that 'only those showing significant differences were presented here' after measuring many waveform parameters. This selective reporting, without any multiple-comparison correction, inflates the Type I error for the late-component findings (Area[40:100]ms, WHQN1, and gamma power). The authors should report all tested metrics or apply an appropriate correction.

minor comments (5)

[Table 1] The p-value for t_zc1 should be labeled with the test type and direction; the current presentation is ambiguous because a two-tailed p of 0.0556 is not significant at the conventional 0.05 level, while a one-tailed p would be.
[Abstract and §3.2] The phrase 'delay at the onset of early components' is imprecise; t_zc1 is the first zero-crossing between P0 and N1, not an onset latency. The wording should be adjusted to match the actual metric.
[§2.3] For Patient 8, the stimulation duration '[7.3:5.3]s' appears to be a typo, as the start exceeds the end; please correct or clarify.
[§3.5] The gamma-band comparison is performed on five time windows but reported as a single pooled comparison; the number of comparisons and whether any correction was applied should be stated.
[Figure 3B inset] The term 'phase delay' could be confused with a phase shift in steady-state signals; a more neutral term such as 'latency difference' would be clearer.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the central DCR/ACEP comparisons are new measurements, not reduced to fitted parameters or self-citations.

full rationale

The paper is an empirical comparison rather than a derivation from fitted parameters. The central reported differences (t_zc1 delay, WN1, WHQN1, Area[40:100]ms, slope minima, and gamma power) are measured from new ECoG data and compared with paired t-tests; no equation reconstructs these outcomes from inputs or from the authors' prior papers. The conduction-velocity estimate in Section 3.2 is a post-hoc arithmetic combination of measured delays and Euclidean distances, not a fitted parameter used to predict the delay, so the delay does not reduce to the estimate. Self-citations (Rossel et al. 2023; Schlosser-Perrin et al. 2023; Boyer et al. 2021b) provide background waveform conventions and prior observations, but the load-bearing delay and shape comparisons do not depend on any self-cited uniqueness theorem or ansatz. The one-tailed interpretation of t_zc1, the selective reporting of metrics that reached significance, and the distance-proxy caveats are statistical and interpretive weaknesses rather than circularity. No step in the paper's argument is equivalent to its own inputs by construction.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The analysis rests on standard neurophysiological assumptions about DES activation order and the canonical EP waveform, plus a geometric assumption that Euclidean surface distance approximates conduction path length. The free parameters are thresholds and time windows chosen in the analysis, some reported only after seeing which tests came out significant.

free parameters (3)

Response amplitude inclusion threshold = 100 µV
Responses with amplitudes below 100 µV were excluded from the comparison (Section 2.3). This threshold is stated but not justified, and could differently affect the DCR and ACEP groups.
Late-component time windows = 40-100 ms for area; 50-80 ms for slope
The time windows for Area and Min(d/dt) metrics were chosen in the analysis, and Section 2.4 states that only metrics showing significant differences were presented. This makes the selected windows a potential post-hoc choice.
Gamma frequency and window = 50 Hz; 20-40 ms pooled
The gamma activity metric focuses on a single frequency (50 Hz) and time windows that were pooled after observing the time-frequency maps (Section 2.5).

assumptions (5)

domain assumption DES initially activates larger myelinated axons before smaller or unmyelinated elements, and conduction velocity scales with fiber diameter (v = 6 x D).
Used in the Introduction and Discussion to predict ACEP delays and to interpret P0 as a synchronized action potential volley (Section 1).
domain assumption DCR, ACEP, and CCEP share a canonical P0-N1-after-positivity waveform produced by the same electrogenic mechanisms at the recording site.
Essential for comparing t_zc1 and other features across response types; asserted in the Discussion, with support from prior literature but not independently proven in this study.
domain assumption The first zero-crossing t_zc1 marks the onset of the N1 component and is comparable between DCR and ACEP despite amplitude and filtering differences.
Used as the primary delay metric (Figure 3 inset, Table 1); if amplitude shifts the zero-crossing differently for the two response types, the delay estimate is biased.
domain assumption The 3D Euclidean distance between the DES site and the recording electrode approximates the axonal path length traveled by the evoked volley.
Used in Section 3.2 to convert measured delays into conduction velocities; the authors acknowledge that actual fiber paths are longer and that very short ACEPs may involve direct cortical stimulation from below.
standard math Paired t-tests on per-patient averaged metrics are appropriate for n=9 patients.
Statistical method in Section 2.6; normality was checked, but the analysis does not account for the varying number of trains per patient or multiple metrics.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Shapes of direct cortical responses vs. short-range axono-cortical evoked potentials: The effects of direct electrical stimulation applied to the human brain." pith.science (2026). https://pith.science/paper/M4BFE3SR

@misc{pith2026241116360,
  author       = {Pith},
  title        = {Pith review of: Shapes of direct cortical responses vs. short-range axono-cortical evoked potentials: The effects of direct electrical stimulation applied to the human brain},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/M4BFE3SR}},
  note         = {Machine review of arXiv:2411.16360}
}

abstract

Objective: Direct cortical responses (DCR) and axono-cortical evoked potentials (ACEP) are generated by electrically stimulating the cortex either directly or indirectly through white matter pathways, potentially leading to different electrogenic processes. For ACEP, the slow conduction velocity of axons (median around 4 m.s$^{-1}$) is anticipated to induce a delay. For DCR, direct electrical stimulation (DES) of the cortex is expected to elicit additional cortical activity involving smaller and slower non-myelinated axons. We tried to validate these hypotheses. Methods: DES was administered either directly on the cortex or to white matter fascicles within the resection cavity, while recording DCR or ACEP at the cortical level in nine patients. Results: Short but significant delays (around 2 ms) were measurable for ACEP immediately following the initial component (around 7 ms). Subsequent activities (around 40 ms) exhibited notable differences between DCR and ACEP, suggesting the presence of additional cortical activities for DCR. Conclusion: Distinctions between ACEPs and DCRs can be made based on a delay at the onset of early components and the dissimilarity in the shape of the later components >40 ms after the DES artifact). Significance: The comparison of different types of evoked potentials allows to better understand the effects of DES.

Figures

Figures reproduced from arXiv: 2411.16360 by the authors.

**Figure 4.** (A) EP Waveforms illustrating the canonical form observed in recorded signals, with N1 and P0 as principal components elicited by stimulation followed by signal renormalization upon return to spontaneous activity. (B). Explanation of waveform variations due to the combination of distinct macroscopic dipoles or generators with different orientations. (1) Brain surgery case: the stimulation applied in the cavity (show… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

4 extracted references · 4 canonical work pages

[161]

Responses of cortex to direct electrical stimuli applied at different depths

https://doi.org/10.1113/jphysiol.1936.sp003427 Bishop GH, Clare MH. Responses of cortex to direct electrical stimuli applied at different depths. Journal of Neurophysiology 1953;16:1–19. https://doi.org/10.1152/jn.1953.16.1.1 Blair EA, Erlanger J. A comparison of the characteristics of axons trough their individual electrical responses. American Journal o...

work page arXiv 1936
[219]

The conduction properties of axons in central white *, Progress in Neurobiology 1977

https://doi.org/10.1038/newbio238217a0 26 Waxman SG, Swadlow HA. The conduction properties of axons in central white *, Progress in Neurobiology 1977;. Pergamon Press Yamao Y , Matsumoto R, Kunieda T, Arakawa Y , Kobayashi K, Usami K, et al. Intraoperative dorsal language network mapping by using single-pulse electrical stimulation. Human Brain Mapping 20...

work page doi:10.1038/newbio238217a0 1977
[1890]

Measuring the electrophysiological effects of direct electrical stimulation after awake brain surgery, in: Journal of Neural Engineering 2020

https://doi.org/10.1016/j.clinph.2017.07.402 Vincent MA, Bonnetblanc F, Mandonnet E, Boyer A, Duffau H, Guiraud D. Measuring the electrophysiological effects of direct electrical stimulation after awake brain surgery, in: Journal of Neural Engineering 2020;. Institute of Physics Publishing. https://doi.org/10.1088/1741-2552/ab5cdd Waxman SG, Bennett MVL. ...

work page doi:10.1016/j.clinph.2017.07.402 2017
[4729]

Electrophysiological brain mapping: Basics of recording evoked potentials induced by electrical stimulation and its physiological spreading in the human brain

https://doi.org/10.1002/hbm.22948 Vincent M, Guiraud D, Duffau H, Mandonnet E, Bonnetblanc F. Electrophysiological brain mapping: Basics of recording evoked potentials induced by electrical stimulation and its physiological spreading in the human brain. Clinical Neurophysiology 2017;128:1886–

work page doi:10.1002/hbm.22948

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [161]

Responses of cortex to direct electrical stimuli applied at different depths

https://doi.org/10.1113/jphysiol.1936.sp003427 Bishop GH, Clare MH. Responses of cortex to direct electrical stimuli applied at different depths. Journal of Neurophysiology 1953;16:1–19. https://doi.org/10.1152/jn.1953.16.1.1 Blair EA, Erlanger J. A comparison of the characteristics of axons trough their individual electrical responses. American Journal o...

work page arXiv 1936

[2] [219]

The conduction properties of axons in central white *, Progress in Neurobiology 1977

https://doi.org/10.1038/newbio238217a0 26 Waxman SG, Swadlow HA. The conduction properties of axons in central white *, Progress in Neurobiology 1977;. Pergamon Press Yamao Y , Matsumoto R, Kunieda T, Arakawa Y , Kobayashi K, Usami K, et al. Intraoperative dorsal language network mapping by using single-pulse electrical stimulation. Human Brain Mapping 20...

work page doi:10.1038/newbio238217a0 1977

[3] [1890]

Measuring the electrophysiological effects of direct electrical stimulation after awake brain surgery, in: Journal of Neural Engineering 2020

https://doi.org/10.1016/j.clinph.2017.07.402 Vincent MA, Bonnetblanc F, Mandonnet E, Boyer A, Duffau H, Guiraud D. Measuring the electrophysiological effects of direct electrical stimulation after awake brain surgery, in: Journal of Neural Engineering 2020;. Institute of Physics Publishing. https://doi.org/10.1088/1741-2552/ab5cdd Waxman SG, Bennett MVL. ...

work page doi:10.1016/j.clinph.2017.07.402 2017

[4] [4729]

Electrophysiological brain mapping: Basics of recording evoked potentials induced by electrical stimulation and its physiological spreading in the human brain

https://doi.org/10.1002/hbm.22948 Vincent M, Guiraud D, Duffau H, Mandonnet E, Bonnetblanc F. Electrophysiological brain mapping: Basics of recording evoked potentials induced by electrical stimulation and its physiological spreading in the human brain. Clinical Neurophysiology 2017;128:1886–

work page doi:10.1002/hbm.22948