Auditory cortex cortico-cortical connectivity correlates with neural entrainment to music across motor and affective brain networks
DOI: Pending DOI
Synopsis
Music engages sensory-driven auditory processing and predictive expectations about a song's temporal structure. These processes involve sensory, motor, and language areas, but how connectivity measured with cortico-cortical evoked potentials (CCEPs) differentially shape sensory-driven and predictive processing remains unknown. Using stereoEEG recordings during music listening and CCEPs, we show that effective connectivity is associated with musical beat-locked changes in neural activity, with a larger effect for a temporally regular metronome compared to naturalistic music. This effect is strongest in motor regions while associative cortices show a larger effect for complex music. These results suggest that effective connectivity may support top-down and bottom-up auditory processing.
Background
Accurate perception of the timing of sounds is central to several human behaviors including speech, music, and dancing. Rhythmic sounds synchronize neural activity in both humans and non-human primates1–3, with spectral power increases at the stimulus frequency while phase coherence decreases ~200ms before beats, reflecting anticipatory excitability increases4. Recent intracranial studies have shown that complex acoustic information and rhythms are represented in high-frequency broadband (HFB) activity (80-200 Hz), a proxy measure for neuronal firing5–7, in the superior temporal gyrus (STG or auditory cortex)2 and across the cortex8–10. These signals reflect sensory-driven processing of acoustic temporal information alongside predictive motor planning, with temporal prediction occurring even in the absence of an auditory stimulus11,12. Language studies also show feedforward and feedback activity throughout auditory pathways during speech comprehension and production13. Cortico-cortical evoked potentials (CCEPs), time-locked neural responses to single pulse direct electrical stimulation, have previously been used in humans to study connectivity between regions during auditory processing and speech planning10,14–16. Yet how auditory cortex effective connectivity constrains auditory temporal prediction across the brain and whether this differs for bottom-up and top-down functional pathways remains unknown.
Methods
We recorded stereo-EEG (sEEG) from 5 neurosurgical patients (Fig. 1A) as they listened to a metronome and naturalistic music. Music was selected from the DEAM database17 based on tempo (~120 beats per minute), and spanned multiple genres. Preprocessed sEEG data were bipolar re-referenced, and we extracted HFB power using morlet wavelet convolution18. Musical beats were detected from audio files, and neural data was extracted +-250ms around each music beat and baseline corrected to the mean of the beat window. Maximum HFB power was extracted on the averaged post-beat time period (0-250ms) separately for metronome and complex music beats for each channel. Single pulse stimulation was delivered to STG in 4 patients (6 or 4 mA, biphasic, 0.2ms pulse width, 0.5 Hz) for 44 pulse events, enabling us to assess electrically evoked activity in auditory pathways. Resulting CCEPs were bipolar re-referenced, baseline corrected to the pre-pulse mean, and averaged (Fig. 1B). Artifact was removed for plotting purposes. Peak-to-peak (p2p) metrics were computed in the 10-50 ms window post stimulation pulse to capture the N1 component, a standard index of effective connectivity strength in prior CCEP studies15, and averaged across STG stimulation sites to yield one connectivity estimate per channel. We employed a linear mixed effects model to assess whether CCEPs during STG stimulation predicted beat-locked HFB power and differed between conditions (metronome vs. music). Regional correlations were computed using Spearman’s correlation for regions with a minimum of 2 patients and 8 channels.
Results
Figure 1: A) Electrode coverage across participants plotted in standard MNI space. B) One example of a preprocessed CCEP during STG stimulation from one neighboring channel. Grey box indicates the artifact period (0-10ms) post stimulation. C) One STG channel’s averaged HFB activity in the beat window for metronome (top) and complex music (bottom) conditions as heat maps (left) and time series averaged across HFB frequencies (right). D) P2p predicts post-beat HFB power, with a larger effect in the metronome condition (teal) compared to complex music (orange). Data points represent individual channels across subjects for each condition. Shaded regions represent 95% confidence intervals. (E) Regional Spearman’s correlation values for the (left) top 3 regions with the greatest positive (metronome>music) and STG, and (right) negative (music>metronome) difference between conditions. N’s represent channel counts across participants.
First, we identified bipolar channels that exhibited HFB synchronization to musical beats regardless of the musical stimulus. We hypothesized that the majority of significant channels would be in auditory cortex (STG) and motor planning areas (Inferior Frontal Gyrus/IFG, Operculum/Oper), as suggested by prior literature3,10,19. Using cluster based permutation testing, we observed a significant change in HFB power locked to the musical beat across sensory and motor regions (Fig. 1C). 82% of channels in STG, 33% in Fusiform Gyrus (FG), 22% in Operculum, 22% in Supramarginal Gyrus (SMG), 16% in IFG, and 25% in Inferior Temporal Gyrus (ITG) and Superior Temporal Sulcus (STS) exhibited modulation in their HFB power in the beat-window. Next, we assessed whether effective connectivity predicted musical beat-evoked HFB power. Given prior literature demonstrating that beat processing engages sensory and motor areas3, and studies showing that these regions are also effectively connected at rest10, we hypothesized that CCEP p2p amplitude during STG stimulation would correlate with music-induced changes in HFB power, and that this effect would be largest for the metronome condition where musical beats are more temporally isolated and regular. We also hypothesized that there would be a dissociation between sensory motor regions (Operculum, STG, PCG) and higher order cortices involved in emotion processing such as Orbitofrontal regions (OrbG), where sensory-motor regions show a greater correlation between beat-locked HFB power and CCEP p2p for the metronome while associative regions show a greater correlation for complex music.
We ran a linear mixed effects model to assess the relationship between the p2p from 10-50ms post CCEP stimulation to STG to the peak post-beat HFB power. We demonstrate a significant main effect of the p2p on HFB power (β=0.13, p<0.001), demonstrating that neural entrainment to musical beats across the brain may be constrained by effective connectivity. We also reveal a significant interaction with condition (Fig. 1D), with greater predictability for the metronome beat-locked HFB activity, supporting our hypothesis (β=0.14, p<0.001). With limited spatial coverage and 4 subjects, preliminary investigation into regional differences provides initial support for our hypothesis, with PCG and operculum, nodes of the motor network, showing the first and third largest differences in correlation between metronome and music (Fig. 1E, left). OrbG, medial frontal (MFC), and IFG channels showed the largest effect in the opposite direction, with larger correlations for the music condition (Fig. 1E, right).
Discussion
Our finding that effective connectivity predicted beat-locked HFB more strongly for temporally regular rhythms in motor planning regions suggests that cortico-cortical pathways from motor to auditory cortex serve as relevant infrastructure for top-down temporal prediction, consistent with prior proposals that motor regions generate and transmit beat timing signals to auditory cortex. The opposite pattern in OrbG and MFC suggests that connectivity through affective pathways is potentially recruited by the acoustic and emotional complexity of naturalistic music20–22. These regional differences provide initial evidence for a network-level dissociation in how cortico-cortical connectivity supports auditory temporal prediction. These results support our hypothesis that effective connectivity differentially constrains sensory-driven and predictive auditory activity, and suggest that effective connectivity is related to the functional organization of the auditory hierarchy from low-level sensory processing to higher-order affective and predictive processing.
Future work will examine CCEP latencies, including delayed N2 responses (50-250ms) across motor planning and affective regions and assess how those relate to the type of stimulus. Additionally, as ongoing data collection increases sample sizes, future analyses will assess how emotional responses to music modulate the relationship between effective connectivity and beat-locked neural activity, and test whether effective connectivity between auditory-motor and affective regions predicts emotional responses to musical stimuli.
Disclosures
CJK holds equity in Alto Neuroscience, Inc. and Flow Neuroscience, Inc.
References
1. Ross, J. M., Comstock, D. C., Iversen, J. R., Makeig, S. & Balasubramaniam, R. Cortical mu rhythms during action and passive music listening. J. Neurophysiol. 127, 213–224 (2022).
2. Nozaradan, S., Peretz, I., Missal, M. & Mouraux, A. Tagging the Neuronal Entrainment to Beat and Meter. J. Neurosci. 31, 10234–10240 (2011).
3. Nozaradan, S. et al. Intracerebral evidence of rhythm transform in the human auditory cortex. Brain Struct. Funct. 222, 2389–2404 (2017).
4. Ross, J. M. et al. Sensory Entrained TMS (seTMS) Enhances Motor Cortex Excitability. Hum. Brain Mapp. 46, e70267 (2025).
5. Lachaux, J.-P., Axmacher, N., Mormann, F., Halgren, E. & Crone, N. E. High-frequency neural activity and human cognition: Past, present and possible future of intracranial EEG research. Prog. Neurobiol. 98, 279–301 (2012).
6. Leszczyński, M. et al. Dissociation of broadband high-frequency activity and neuronal firing in the neocortex. Sci. Adv. 6, eabb0977 (2020).
7. Ray, S., Crone, N. E., Niebur, E., Franaszczuk, P. J. & Hsiao, S. S. Neural Correlates of High-Gamma Oscillations (60–200 Hz) in Macaque Local Field Potentials and Their Potential Implications in Electrocorticography. J. Neurosci. 28, 11526–11536 (2008).
8. Herff, S. A. et al. Prefrontal High Gamma in ECoG Tags Periodicity of Musical Rhythms in Perception and Imagination. eneuro 7, ENEURO.0413-19.2020 (2020).
9. Sturm, I., Blankertz, B., Potes, C., Schalk, G. & Curio, G. ECoG high gamma activity reveals distinct cortical representations of lyrics passages, harmonic and timbre-related changes in a rock song. Front. Hum. Neurosci. 8, (2014).
10. Nishida, M. et al. Brain network dynamics in the human articulatory loop. Clin. Neurophysiol. 128, 1473–1487 (2017).
11. Fujioka, T., Trainor, L. J., Large, E. W. & Ross, B. Internalized Timing of Isochronous Sounds Is Represented in Neuromagnetic Beta Oscillations. J. Neurosci. 32, 1791–1802 (2012).
12. Nozaradan, S., Peretz, I. & Keller, P. E. Individual Differences in Rhythmic Cortical Entrainment Correlate with Predictive Behavior in Sensorimotor Synchronization. Sci. Rep. 6, 20612 (2016).
13. Wang, R. et al. Distributed feedforward and feedback cortical processing supports human speech production. Proc. Natl. Acad. Sci. 120, e2300255120 (2023).
14. Barczak, A. et al. Top-down, contextual entrainment of neuronal oscillations in the auditory thalamocortical circuit. Proc. Natl. Acad. Sci. 115, (2018).
15. Matsumoto, R. et al. Functional connectivity in the human language system: a cortico-cortical evoked potential study. Brain 127, 2316–2330 (2004).
16. Wang, Y. et al. Spatial-Temporal Functional Mapping Combined With Cortico-Cortical Evoked Potentials in Predicting Cortical Stimulation Results. Front. Hum. Neurosci. 15, 661976 (2021).
17. Aljanaki, A., Yang, Y.-H. & Soleymani, M. Developing a benchmark for emotional analysis of music. PLOS ONE 12, e0173392 (2017).
18. Cohen, M. X. Analyzing Neural Time Series Data: Theory and Practice. (The MIT Press, 2014). doi:10.7551/mitpress/9609.001.0001.
19. Bellier, L. et al. Music can be reconstructed from human auditory cortex activity using nonlinear decoding models. PLOS Biol. 21, e3002176 (2023).
20. Omigie, D. et al. An Intracranial EEG Study of the Neural Dynamics of Musical Valence Processing. Cereb. Cortex 25, 4038–4047 (2015).
21. Blood, A. J. & Zatorre, R. J. Intensely pleasurable responses to music correlate with activity in brain regions implicated in reward and emotion. Proc. Natl. Acad. Sci. 98, 11818–11823 (2001).
22. Mas-Herrero, E., Dagher, A. & Zatorre, R. J. Modulating musical reward sensitivity up and down with transcranial magnetic stimulation. Nat. Hum. Behav. 2, 27–32 (2017).