Re-evaluating the contribution of sensorimotor mu rhythm phase and power to human corticospinal output: A replication study
Tharan Suresh, Sara J. Hussain
Abstract
Transcranial magnetic stimulation (TMS) is a noninvasive brain stimulation technique that can be used to both probe and modulate human corticospinal tract function. Conventionally, TMS is delivered to the primary motor cortex (M1) irrespective of ongoing sensorimotor rhythm activity. Such brain state-independent methods of TMS delivery may partially explain the wide response variability to TMS interventions targeting the corticospinal tract. Accordingly, recent studies using real-time brain state-dependent TMS [[1]Wischnewski M. Haigh Z.J. Shirinpour S. Alekseichuk I. Opitz A. The phase of sensorimotor mu and beta oscillations has the opposite effect on corticospinal excitability.Brain Stimul: Basic, Translational, and Clinical Research in Neuromodulation. 2022; 15: 1093-1100https://doi.org/10.1016/j.brs.2022.08.005Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar,[2]Zrenner C. Desideri D. Belardinelli P. Ziemann U. Real-time EEG-defined excitability states determine efficacy of TMS-induced plasticity in human motor cortex.Brain Stimul. 2018; 11: 374-389https://doi.org/10.1016/j.brs.2017.11.016Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar] or offline trial sorting approaches [[3]Hussain S.J. Claudino L. Bönstrup M. Norato G. Cruciani G. Thompson R. Zrenner C. Ziemann U. Buch E. Cohen L.G. Sensorimotor oscillatory phase–power interaction gates resting human corticospinal output.Cerebr Cortex. 2019; 29: 3766-3777https://doi.org/10.1093/cercor/bhy255Crossref PubMed Scopus (32) Google Scholar,[4]Ozdemir R.A. Kirkman S. Magnuson J.R. Fried P.J. Pascual-Leone A. Shafi M.M. Phase matters when there is power: phasic modulation of corticospinal excitability occurs at high amplitude sensorimotor mu-oscillations.Neuroimage: Report. 2022; 2100132https://doi.org/10.1016/j.ynirp.2022.100132Crossref PubMed Scopus (1) Google Scholar] showed that corticospinal tract output varies significantly with the sensorimotor rhythmic brain state at the time of TMS delivery. Most work in this domain has focused on the sensorimotor mu rhythm, which oscillates between 8 and 13 Hz and is the spectrally dominant rhythm recorded over sensorimotor cortical regions. For example, single-neuron M1 spiking rates [[5]Haegens S. Nácher V. Luna R. Romo R. Jensen O. α-Oscillations in the monkey sensorimotor network influence discrimination performance by rhythmical inhibition of neuronal spiking.Proc Natl Acad Sci USA. 2011; 108: 19377-19382https://doi.org/10.1073/pnas.1117190108Crossref PubMed Scopus (470) Google Scholar], corticospinal output [[3]Hussain S.J. Claudino L. Bönstrup M. Norato G. Cruciani G. Thompson R. Zrenner C. Ziemann U. Buch E. Cohen L.G. Sensorimotor oscillatory phase–power interaction gates resting human corticospinal output.Cerebr Cortex. 2019; 29: 3766-3777https://doi.org/10.1093/cercor/bhy255Crossref PubMed Scopus (32) Google Scholar,[6]Bergmann T.O. Lieb A. Zrenner C. Ziemann U. Pulsed facilitation of corticospinal excitability by the sensorimotor μ-alpha rhythm.J Neurosci. 2019; 39: 10034-10043https://doi.org/10.1523/JNEUROSCI.1730-19.2019Crossref PubMed Scopus (44) Google Scholar], and interhemispheric communication between homologous M1 subregions [[7]Stefanou M.-I. Desideri D. Belardinelli P. Zrenner C. Ziemann U. Phase synchronicity of μ-rhythm determines efficacy of interhemispheric communication between human motor cortices.J Neurosci. 2018; 38: 10525-10534https://doi.org/10.1523/JNEUROSCI.1470-18.2018Crossref PubMed Scopus (31) Google Scholar] are all increased during sensorimotor mu rhythm trough (i.e., negative peak) phases compared to mu rhythm peak (i.e., positive peak) phases. Further, TMS interventions applied to M1 during mu trough phases preferentially enhance corticospinal transmission and boost motor learning, while identical interventions applied during mu peak phases weakly depress corticospinal transmission and do not influence motor learning [[2]Zrenner C. Desideri D. Belardinelli P. Ziemann U. Real-time EEG-defined excitability states determine efficacy of TMS-induced plasticity in human motor cortex.Brain Stimul. 2018; 11: 374-389https://doi.org/10.1016/j.brs.2017.11.016Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar,[8]Baur D. Galevska D. Hussain S. Cohen L.G. Ziemann U. Zrenner C. Induction of LTD-like corticospinal plasticity by low-frequency rTMS depends on pre-stimulus phase of sensorimotor μ-rhythm.Brain Stimul. 2020; 13: 1580-1587https://doi.org/10.1016/j.brs.2020.09.005Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar,[9]Hussain S.J. Vollmer M.K. Stimely J. Norato G. Zrenner C. Ziemann U. Buch E.R. Cohen L.G. Phase-dependent offline enhancement of human motor memory.Brain Stimul. 2021; 14: 873-883https://doi.org/10.1016/j.brs.2021.05.009Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar]. Recent work has repeatedly shown that the sensorimotor mu rhythm shapes corticospinal output, but the precise mu rhythm features contributing to this effect are still unclear. Some previous studies have identified relationships between mu phase and corticospinal output that do not depend on mu power [[1]Wischnewski M. Haigh Z.J. Shirinpour S. Alekseichuk I. Opitz A. The phase of sensorimotor mu and beta oscillations has the opposite effect on corticospinal excitability.Brain Stimul: Basic, Translational, and Clinical Research in Neuromodulation. 2022; 15: 1093-1100https://doi.org/10.1016/j.brs.2022.08.005Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar,[2]Zrenner C. Desideri D. Belardinelli P. Ziemann U. Real-time EEG-defined excitability states determine efficacy of TMS-induced plasticity in human motor cortex.Brain Stimul. 2018; 11: 374-389https://doi.org/10.1016/j.brs.2017.11.016Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar], while others found no relationship between mu phase and corticospinal output [[10]Madsen K.H. Karabanov A.N. Krohne L.G. Safeldt M.G. Tomasevic L. Siebner H.R. No trace of phase: corticomotor excitability is not tuned by phase of pericentral mu-rhythm.Brain Stimul. 2019; 12: 1261-1270https://doi.org/10.1016/j.brs.2019.05.005Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar]. In contrast, our own work [[3]Hussain S.J. Claudino L. Bönstrup M. Norato G. Cruciani G. Thompson R. Zrenner C. Ziemann U. Buch E. Cohen L.G. Sensorimotor oscillatory phase–power interaction gates resting human corticospinal output.Cerebr Cortex. 2019; 29: 3766-3777https://doi.org/10.1093/cercor/bhy255Crossref PubMed Scopus (32) Google Scholar] and that of others [[4]Ozdemir R.A. Kirkman S. Magnuson J.R. Fried P.J. Pascual-Leone A. Shafi M.M. Phase matters when there is power: phasic modulation of corticospinal excitability occurs at high amplitude sensorimotor mu-oscillations.Neuroimage: Report. 2022; 2100132https://doi.org/10.1016/j.ynirp.2022.100132Crossref PubMed Scopus (1) Google Scholar] found that mu rhythm phase and power interact to influence corticospinal output. Yet, nearly all previous studies, including our own, failed to quantitatively dissociate mu rhythm power into its periodic and aperiodic components. Historically, aperiodic components of neural power spectra (i.e., 1/f trends) have been ignored or considered as noise. Yet, changes in spectral power can be caused by alterations in both periodic and aperiodic components of neural power spectra [[11]Donoghue T. Haller M. Peterson E.J. Varma P. Sebastian P. Gao R. Noto T. Lara A.H. Wallis J.D. Knight R.T. Shestyuk A. Voytek B. Parameterizing neural power spectra into periodic and aperiodic components.Nat Neurosci. 2020; 23 (Article 12)https://doi.org/10.1038/s41593-020-00744-xCrossref PubMed Scopus (343) Google Scholar], which are thought to be generated by distinct neural mechanisms. Here, we aimed to systematically re-evaluate the contribution of sensorimotor mu rhythm phase, mu rhythm power (including both periodic and aperiodic components), and their interaction to corticospinal tract output using a quantitative model-building approach applied to a new single-pulse real-time brain state-dependent TMS dataset acquired from 22 healthy humans. Real-time EEG analysis accurately identified and targeted mu peak (target phase = 90°, observed mean phase = 89.19° [95% CI = 84.12°–94.25°], V = 301.61, p < 0.0001) and mu trough phases (target phase = 270°, observed mean phase = 266.33° [95% CI = 261.55°–271.10°], V = 40.17, p < 0.008, see Fig. 1a-d). As expected, targeted phases differed significantly between peak and trough conditions (F = 2144.5, p < 0.0001). For random phases, we observed a small deviation from circular uniformity in the 20° direction (R = 3.01, p = 0.04, observed mean phase = 20.51° [95% CI = −32.43° – 73.44°]). To systematically analyze the influence of mu phase, mu power, and their interaction on corticospinal output, we fit separate linear-mixed effects models to trial-by-trial log-transformed MEP amplitudes. All models included random effects of subject, but each model differed in their included fixed effects (see Supplementary Table 1). Model 1 included a fixed effect of mu phase; this model revealed a significant effect of mu phase (F = 3.12, p = 0.04). Model 2 included fixed effects of phase, mixed mu power (i.e., the arithmetic sum of periodic and aperiodic mu power components), and their interaction. Model 2 revealed main effects of phase (F = 15.23, p < 0.001) and mixed mu power (F = 25.35, p < 0.001), as well as their interaction (F = 18.3, p < 0.001). Model 3 included fixed effects of phase, the periodic mu power component, and their interaction. Model 3 revealed main effects of the periodic mu power component (F = 5.14; p = 0.02), phase (F = 6.76, p = 0.001), and their interaction (F = 9.52, p = 0.001). Model 4 included fixed effects of phase, the aperiodic mu power component, and their interaction. For model 4, there was no main effect of the aperiodic mu component (F = 3.61, p = 0.06), but a significant main effect of phase (F = 8.18, p < 0.001), and a significant interaction between the aperiodic mu power component and phase interaction (F = 7.18, p < 0.001). See Supplementary Table 2 for full statistical results. Quantitative model comparison based on Akaike and Bayesian Information Criteria values showed that model 2 best fit the data (see Supplementary Table 1). For model 2, the slope between mixed mu power and log-transformed MEP amplitudes was higher for mu trough than for mu peak phases (t = 5.96, uncorrected and Bonferroni-corrected p < 0.001) and lower for mu peak than for random phases (t = 3.82, uncorrected and Bonferroni-corrected p < 0.001). The slope between mixed mu power and log-transformed MEP amplitudes was also higher for mu trough than random phases (t = 2.05, uncorrected p = 0.04, see Fig. 1F), but this difference did not survive correction for multiple comparisons (Bonferroni-corrected p = 0.122). Our results demonstrate that sensorimotor mu rhythm phase and mu rhythm power interdependently shape corticospinal output. Specifically, corticospinal output was better explained by a statistical model that accounted for the interaction between mu rhythm phase and mixed mu power than by statistical models that accounted for mu rhythm phase only or the interaction between mu rhythm phase and periodic or aperiodic mu power components. These findings highlight the complex interplay between sensorimotor mu rhythm phase and power in shaping corticospinal tract activity, emphasizing the importance of considering both mu power and phase metrics when using real-time phase-dependent TMS for either measurement or interventional purposes. Our findings are consistent with previous studies reporting that phase-power interactions shape corticospinal output [[3]Hussain S.J. Claudino L. Bönstrup M. Norato G. Cruciani G. Thompson R. Zrenner C. Ziemann U. Buch E. Cohen L.G. Sensorimotor oscillatory phase–power interaction gates resting human corticospinal output.Cerebr Cortex. 2019; 29: 3766-3777https://doi.org/10.1093/cercor/bhy255Crossref PubMed Scopus (32) Google Scholar,[4]Ozdemir R.A. Kirkman S. Magnuson J.R. Fried P.J. Pascual-Leone A. Shafi M.M. Phase matters when there is power: phasic modulation of corticospinal excitability occurs at high amplitude sensorimotor mu-oscillations.Neuroimage: Report. 2022; 2100132https://doi.org/10.1016/j.ynirp.2022.100132Crossref PubMed Scopus (1) Google Scholar] but conflict with others that reported no such interactions [[1]Wischnewski M. Haigh Z.J. Shirinpour S. Alekseichuk I. Opitz A. The phase of sensorimotor mu and beta oscillations has the opposite effect on corticospinal excitability.Brain Stimul: Basic, Translational, and Clinical Research in Neuromodulation. 2022; 15: 1093-1100https://doi.org/10.1016/j.brs.2022.08.005Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar,[2]Zrenner C. Desideri D. Belardinelli P. Ziemann U. Real-time EEG-defined excitability states determine efficacy of TMS-induced plasticity in human motor cortex.Brain Stimul. 2018; 11: 374-389https://doi.org/10.1016/j.brs.2017.11.016Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar,[6]Bergmann T.O. Lieb A. Zrenner C. Ziemann U. Pulsed facilitation of corticospinal excitability by the sensorimotor μ-alpha rhythm.J Neurosci. 2019; 39: 10034-10043https://doi.org/10.1523/JNEUROSCI.1730-19.2019Crossref PubMed Scopus (44) Google Scholar]. Methodological differences in choice of real-time EEG analysis algorithms, pulse type, inter-stimulus interval (ISI), experimental design, window length for power calculation, number of trials and participant pre-selection based on electrophysiological characteristics may contribute to these differential results (see Supplementary Table 3 for summary of methodological differences across studies). While the effect of mu rhythm phase on corticospinal output has been repeatedly demonstrated, these studies often use relatively short (i.e., 2–3 s) ISIs. It is therefore crucial to consider the potential confounding effects of ISI and phase-dependent entrainment [[12]Thut G. Veniero D. Romei V. Miniussi C. Schyns P. Gross J. Rhythmic TMS causes local entrainment of natural oscillatory signatures.Curr Biol. 2011; 21: 1176-1185https://doi.org/10.1016/j.cub.2011.05.049Abstract Full Text Full Text PDF PubMed Scopus (364) Google Scholar] when interpreting their results. For example, repeatedly delivering mu phase-dependent single-pulse TMS every 2–3 s could progressively entrain the sensorimotor rhythm, increasing its power and exaggerating any mu phase-dependency of corticospinal output. Here, we delivered mu phase-dependent single-pulse TMS every ∼5 s, rather than every ∼2–3 s as done previously [[1]Wischnewski M. Haigh Z.J. Shirinpour S. Alekseichuk I. Opitz A. The phase of sensorimotor mu and beta oscillations has the opposite effect on corticospinal excitability.Brain Stimul: Basic, Translational, and Clinical Research in Neuromodulation. 2022; 15: 1093-1100https://doi.org/10.1016/j.brs.2022.08.005Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar,[2]Zrenner C. Desideri D. Belardinelli P. Ziemann U. Real-time EEG-defined excitability states determine efficacy of TMS-induced plasticity in human motor cortex.Brain Stimul. 2018; 11: 374-389https://doi.org/10.1016/j.brs.2017.11.016Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar,[8]Baur D. Galevska D. Hussain S. Cohen L.G. Ziemann U. Zrenner C. Induction of LTD-like corticospinal plasticity by low-frequency rTMS depends on pre-stimulus phase of sensorimotor μ-rhythm.Brain Stimul. 2020; 13: 1580-1587https://doi.org/10.1016/j.brs.2020.09.005Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar]. In addition to reducing the likelihood of entraining the sensorimotor mu rhythm, our longer inter-stimulus interval is also less likely to induce cumulative LTP-like plasticity effects that may elicit larger trough-specific MEP amplitudes than would be expected in the absence of such plasticity (see Ref. [[2]Zrenner C. Desideri D. Belardinelli P. Ziemann U. Real-time EEG-defined excitability states determine efficacy of TMS-induced plasticity in human motor cortex.Brain Stimul. 2018; 11: 374-389https://doi.org/10.1016/j.brs.2017.11.016Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar]). In addition, some previous studies did not explicitly test for phase-power interaction effects [[2]Zrenner C. Desideri D. Belardinelli P. Ziemann U. Real-time EEG-defined excitability states determine efficacy of TMS-induced plasticity in human motor cortex.Brain Stimul. 2018; 11: 374-389https://doi.org/10.1016/j.brs.2017.11.016Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar,[6]Bergmann T.O. Lieb A. Zrenner C. Ziemann U. Pulsed facilitation of corticospinal excitability by the sensorimotor μ-alpha rhythm.J Neurosci. 2019; 39: 10034-10043https://doi.org/10.1523/JNEUROSCI.1730-19.2019Crossref PubMed Scopus (44) Google Scholar,[8]Baur D. Galevska D. Hussain S. Cohen L.G. Ziemann U. Zrenner C. Induction of LTD-like corticospinal plasticity by low-frequency rTMS depends on pre-stimulus phase of sensorimotor μ-rhythm.Brain Stimul. 2020; 13: 1580-1587https://doi.org/10.1016/j.brs.2020.09.005Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar]. Those that did used different windows for power spectral analysis (ranging from 150 ms to 1 s; [1Wischnewski M. Haigh Z.J. Shirinpour S. Alekseichuk I. Opitz A. The phase of sensorimotor mu and beta oscillations has the opposite effect on corticospinal excitability.Brain Stimul: Basic, Translational, and Clinical Research in Neuromodulation. 2022; 15: 1093-1100https://doi.org/10.1016/j.brs.2022.08.005Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar, 2Zrenner C. Desideri D. Belardinelli P. Ziemann U. Real-time EEG-defined excitability states determine efficacy of TMS-induced plasticity in human motor cortex.Brain Stimul. 2018; 11: 374-389https://doi.org/10.1016/j.brs.2017.11.016Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar, 3Hussain S.J. Claudino L. Bönstrup M. Norato G. Cruciani G. Thompson R. Zrenner C. Ziemann U. Buch E. Cohen L.G. Sensorimotor oscillatory phase–power interaction gates resting human corticospinal output.Cerebr Cortex. 2019; 29: 3766-3777https://doi.org/10.1093/cercor/bhy255Crossref PubMed Scopus (32) Google Scholar]. We suggest using a window length of 500 ms for power spectral analysis, primarily because brain oscillations are nonstationary and 500 ms analysis windows are commonly used in real-time phase targeting algorithms [[2]Zrenner C. Desideri D. Belardinelli P. Ziemann U. Real-time EEG-defined excitability states determine efficacy of TMS-induced plasticity in human motor cortex.Brain Stimul. 2018; 11: 374-389https://doi.org/10.1016/j.brs.2017.11.016Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar,[10]Madsen K.H. Karabanov A.N. Krohne L.G. Safeldt M.G. Tomasevic L. Siebner H.R. No trace of phase: corticomotor excitability is not tuned by phase of pericentral mu-rhythm.Brain Stimul. 2019; 12: 1261-1270https://doi.org/10.1016/j.brs.2019.05.005Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar,[13]Shirinpour S. Alekseichuk I. Mantell K. Opitz A. Experimental evaluation of methods for real-time EEG phase-specific transcranial magnetic stimulation.J Neural Eng. 2020; 17046002https://doi.org/10.1088/1741-2552/ab9dbaCrossref PubMed Scopus (12) Google Scholar]. In conclusion, our results demonstrate that sensorimotor mu rhythm phase and power interdependently shape human corticospinal tract function. Our findings replicate previous reports of mu rhythm phase-power interactions, emphasizing the robustness of this relationship. We therefore suggest that real-time brain state-dependent TMS interventions targeting the motor cortex or corticospinal tract monitor both mu rhythm phase and power. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. SJH is supported by K12HD093427. The following is the Supplementary data to this article. Download .docx (.04 MB) Help with docx files Multimedia component 1