Markus Siegel

Magnetomyography (MMG) holds great potential as a contactless alternative to electromyography (EMG). However, the spatiotemporal dynamics of the myomagnetic field in vivo remain largely unknown, limiting signal interpretability and hence the use of MMG. Here, the myomagnetic field of the abductor digiti minimi muscle is measured in vivo at 69‐by‐2 effective positions by scanning a grid of six triaxial optically pumped magnetometers (OPMs) along the muscle, while reproducible muscle contractions are evoked electrically. In this way, the compound muscle action field (CMAF) is measured with high density (1 mm sampling steps) and its evolution is referenced to the underlying anatomy based on 3D scans and magnetic resonance images. A model is used to simulate the expected signals. It suggests that the clinically relevant innervation zone (IZ) manifests as a zero‐crossing of the azimuthal CMAF component. Based on this, contactless IZ localization is demonstrated in three subjects by projecting the zero‐crossing onto the skin. The estimated IZ localizations deviate by 1.7 mm to 4.1 mm from estimates with simultaneously recorded 64‐channel surface EMG. With the first millimeter‐scale spatiotemporal mapping of magnetic muscle activity in vivo, this study paves the way for the use of MMG in biomedical applications.

Objective . Complexity-based metrics have been applied to surface electromyography (sEMG) to characterize fatigue-related changes in the temporal structure of myoelectric signals beyond amplitude and spectral features. Optically pumped magnetometers (OPM) are sensors that enable non-invasive recordings of magnetomyography (MMG) signals from skeletal muscle and are increasingly used to complement sEMG; however, it remains unclear whether complexity measures derived from magnetic recordings are comparable to those obtained from sEMG. Here, we directly compared fatigue-related dynamics of conventional and complexity-based signal features of sEMG and OPM-MMG measured from the biceps brachii during sustained elbow flexion. Approach. Healthy participants performed isometric contractions at 20% maximal voluntary contraction (MVC; 20 min) or 60% MVC (3 min). sEMG and OPM-based MMG were recorded simultaneously, and signal median frequency, root mean square (RMS), and Lempel–Ziv (LZ) complexity were calculated over time. Main results. Across contraction intensities, sEMG and MMG showed consistent fatigue-related changes, characterized by increasing RMS, decreasing median frequency, and a progressive decline in LZ over time. In addition, multiple regression analyses indicated that the decrease in LZ was not fully accounted for by concurrent amplitude or spectral changes, suggesting that complexity captures aspects of signal organization that are not fully explained by established features. Finally, while sEMG showed higher LZ complexity and median frequency at 60% compared to 20% MVC, corresponding intensity-dependent effects were not observed in OPM-based MMG. Significance. These findings suggest that complexity-based metrics capture fatigue-related changes in neuromuscular signal organization beyond conventional measures, and that sEMG and OPM-based MMG provide similar, though modality-specific, information. Together, the results support the use of complexity metrics in multimodal electrophysiological and biomagnetic assessments of neuromuscular fatigue.

The human brain is spatially organized along functional gradients, but the temporal organization of these gradients remains poorly understood. In EEG and MEG, this question has been difficult to address because volume conduction obscures temporal precedence. Here, we introduce lag-specific orthogonalization, an extension of pairwise orthogonalization that allows us to estimate non-zero-lag amplitude-envelope correlations. Applying this approach to large-scale resting-state MEG, we identify electrophysiological lag threads: reproducible maps of temporal ordering that are most coherent in the alpha and beta bands and organize activity from occipital and sensorimotor regions toward frontal and temporal association cortex. These fast electrophysiological lag threads share subject-specific spatial structure with much slower fMRI lag threads, supporting a neural rather than vascular origin of fMRI lag structure. Finally, we show that cortical lag threads are not fixed anatomical constraints. They covary with age and handedness and are selectively reconfigured during cognitive task performance. Together, these findings reveal a frequency-specific electrophysiological temporal hierarchy of the human cortex that links fast neural dynamics to slower hemodynamic propagation structure.

Since young adults hear sounds up to 20 kHz, the loss of extended high-frequency hearing ( EHF ; above 8 kHz) is a hallmark of age-related hearing loss, often progressing from early lifetime. However, this deficit frequently goes undetected because routine clinical hearing tests and most hearing aids are currently limited mostly up to 8 kHz. EHF hearing has been linked to deficits in speech perception in noise and to self-reported hearing. However, it remains elusive how EHF hearing influences speech intelligibility. Here we recorded neuromagnetic brain responses using magnetoencephalography (MEG) within a frequency-tagging speech paradigm designed to probe hierarchical levels of attention and memory-dependent speech processing and recognition. Auditory evoked cortical magnetic field (AEF) responses were significantly reduced in both left and right brain hemispheres in individuals with impaired EHF hearing compared to those subjects with rather preserved EHF hearing. A gradual reinforcement of left-hemispheric AEF seen over age was not observed in young adults (19-29 y) with preserved EHF hearing. This was linked to stronger auditory brainstem responses (ABR), reflecting better neural synchronized auditory responses at stimulus onset. The reinforced left hemispheric dominance in young adults with impaired EHF hearing, in contrast, was linked to lower ABRs. Our findings suggest that sound energy above 8 kHz contributes through its impact on stimulus-onset synchrony to phase locking of oscillations in the auditory cortex to intelligible speech. Together, the results highlight the need to reconsider the neglect of EHF hearing in both audiological assessment and hearing aid design. Significance: We show here that deficits in extended high-frequency (EHF) hearing, up to now neglected in routine clinical audiometry and hearing aid technology, lead to reduced cortical evoked auditory field (AEF) response amplitudes to attended and unattended speech, even at a young age. A gradual increase in reinforced left-hemispheric AEF responses during attended speech does not occur in young people with good EHF hearing; this is linked to better synchronization of neural responses at the onset of sound. This suggests a crucial role of sounds containing energy above 8kHz in minimizing the need for cognitive resources during active listening. Collectively, our results challenge current clinical practices and underscore the need to incorporate EHF hearing into audiological assessment and hearing aid design.

Turning is the most common trigger for freezing of gait in people with Parkinson’s disease, posing a substantial threat due to falls and the risk of traumatic injury. Within the basal ganglia, the subthalamic nucleus plays a central role in subthalamo-spinal motor integration, yet the neuronal processes underlying regular turning and freezing while turning have sparsely been investigated, hampering both mechanistic understanding and therapeutic conceptualization for freezing of gait. We recorded subthalamic local field potentials synchronized to gait kinematics in the medication-OFF and stimulation-Off state in 13 Parkinson’s disease patients with freezing of gait. We analysed spectral modulation and beta burst characteristics across locomotor states including regular turning, freezing while turning, unimpaired walking, as well as the transition periods before initiating turning. We found specific subthalamic activation-deactivation abnormalities both before and during freezing while turning, confined to the low beta range (12–20 Hz), which were associated with kinematic abnormalities (“sequence effect”) preceding freezing while turning. Our burst analyses revealed significant state-dependent differences with shorter bursts occurring in quasi-automatic movement during unimpaired walking or the transition period of regular turns, whereas freezing while turning was characterized by prolonged bursts. Guided by these neuronal and kinematic insights, we trained a generalized linear mixed-effects logistic regression model based on neuronal oscillatory (burst duration) and kinematic features (mean angular velocity, stride length) extracted from both the pre-turn transition and subsequent turning phase. Classification performance was quantified using leave-one-subject-out cross-validation, showing good discriminatory performance (accuracy 70 ± 19%, AU-ROC: 0.84 ± 0.14) in differentiating between regular turning from freezing while turning. Our findings suggest that defective supraspinal-motor integration precedes and further drives freezing while turning. Beyond these pathophysiological insights, our results may inform the conceptualization of future adaptive neuromodulation therapies.

Objective. Transcranial electric stimulation (tES) is a promising technique to non-invasively modulate human brain activity. However, stimulation artifacts in EEG and MEG recordings severely hinder the study of its online effects. Approach. Here, we introduce a new approach to account for these artifacts. The approach rests on two key ideas. First, we focus on interactions of tES with intrinsic brain activity, which are absent for tES artifacts. Second, rather than removing artifacts, we compare composite signals that share similar artifacts but potentially differ in the interaction of tES with intrinsic brain activity. We follow a simple logic: if tES does not interact with intrinsic brain activity, then neural activity during simultaneous sensory stimulation and tES should equal the linear superposition of the neural effects of each applied alone. Any deviation from this prediction provides evidence for a neural interaction. Main results. We tested this approach in a proof-of-principle MEG study, applying 10 Hz transcranial alternating current stimulation (tACS) during rest and during a 10 Hz visual flicker. We compared neural activity during simultaneous stimulation with that predicted by the linear superposition of flicker and tACS alone and found a phase-dependent interaction between tACS and flicker-evoked brain activity. Significance. Our work establishes a novel approach to investigate online effects of tES and suggests a state-dependent interaction of tACS with human brain activity.

Objective: Complexity-based metrics have been applied to surface electromyography (sEMG) to characterize fatigue-related changes in the temporal structure of myoelectric signals beyond amplitude and spectral features. Optically pumped magnetometers (OPM) are sensors that enable non-invasive recordings of magnetomyographic (MMG) signals from skeletal muscle and are increasingly used to complement surface electromyography; however, it remains unclear whether complexity measures derived from magnetic recordings are comparable to those obtained from sEMG. Here, we directly compared fatigue-related dynamics of conventional and complexity-based signal features of sEMG and OPM-MMG measured from the biceps brachii during sustained elbow flexion. Approach: Healthy participants performed isometric contractions at 20% maximal voluntary contraction (MVC; 20 min) or 60% MVC (3 min). sEMG and OPM-based MMG were recorded simultaneously, and signal median frequency, root mean square (RMS), and Lempel-Ziv (LZ) complexity were calculated over time. Main results: Across contraction intensities, sEMG and MMG showed consistent fatigue-related changes, characterized by increasing RMS, decreasing median frequency, and a progressive decline in LZ over time. In addition, multiple regression analyses indicated that the decrease in LZ was not fully accounted for by concurrent amplitude or spectral changes, suggesting that complexity captures aspects of signal organization that are not fully explained by established features. Finally, while sEMG showed higher LZ complexity and median frequency at 60% compared to 20% MVC, corresponding intensity-dependent effects were not observed in OPM-based MMG. Significance: These findings suggest that complexity-based metrics capture fatigue-related changes in neuromuscular signal organization beyond conventional measures, and that sEMG and OPM-based MMG provide similar, though modality-specific, information. Together, the results support the use of complexity metrics in multimodal electrophysiological and biomagnetic assessments of neuromuscular fatigue.

Objective: Focal cortical dysplasia (FCD) causes drug‐resistant epilepsy requiring presurgical evaluation. Invasive electroencephalographic (EEG) studies demonstrate that sleep modulates epileptic activity, including interictal epileptiform discharges (IEDs), fast oscillations (FOs) in the beta (14–40 Hz) and gamma (40–80 Hz) frequency bands, and seizures. This study aimed to quantify sleep‐associated changes in IEDs, FOs, and seizures in FCD patients using noninvasive magnetoencephalography (MEG). Methods: Nineteen patients with FCD were prospectively recruited and underwent simultaneous MEG/EEG recordings lasting 89 ± 19 min during daytime sleep. Sleep stages were classified from the EEG. Beamformer source signals were computed from the MEG signal to enhance sensitivity for visual detection of IEDs, FOs in the beta and gamma frequency bands, and seizures. Magnetic source imaging (MSI) was performed using the Maximum Entropy on the Mean (MEM) method, which is particularly sensitive to the spatial extent of sources, enabling accurate localization of epileptic activity. Results: N1 sleep was reached in 17 of 19 patients and N2 sleep in 14 of 19 patients. Compared to wakefulness, sleep recordings showed significantly higher rates of FOs and seizures (both p < .05), whereas IED rates showed nonsignificant trends. Ten patients demonstrated FOs or seizures, and 12 showed IEDs. MSI of IEDs demonstrated consistent accuracy across vigilance states, with median Euclidean distances of 12.74 mm (interquartile range [IQR] = 22.74) in wake and 8.34 mm (IQR = 27.58) in sleep, and no systematic amplitude or spatial extent changes. Wavelet‐MEM enabled frequency‐specific source imaging, with FOs and seizures localizing concordantly to FCD lesions in five of seven and seven of eight patients, respectively. Significance: Daytime sleep MEG recordings are clinically feasible and significantly enhance the detection of seizures (37% of patients) and FOs compared to wakefulness. Sleep protocols enable noninvasive capture of ictal patterns—the gold standard for epileptogenic zone localization—alongside increased FO rates. These findings support incorporating sleep into standard MEG protocols for presurgical epilepsy evaluation.

Artificial neural networks (ANNs) offer a data-driven approach to reveal brain regional functions without explicit supervision. Here, we demonstrate that an ANN trained to decode visual stimuli from multi-unit spiking activity in monkeys, can not only reconstruct complex and dynamic scenes, but also spontaneously align with canonical cortical visual functions. Without any region-specific functional priors, the model identifies key brain areas associated with shape, color, and motion processing. We provide strong evidence that, despite low train-test dataset correlation at the recording-site level, the ANN-based model is able to learn task-relevant representations embedded at a high-dimensional population level and achieve reliable decoding performance. Furthermore, by inverting the architecture and retraining, we show that the same network can predict region-specific spiking patterns from visual input, indicating a reciprocal relationship between encoding and decoding. These findings shed light on ANN-based visual decoding as a powerful framework for unsupervised functional alignment in neural systems.

Optically pumped magnetometers (OPMs) are compact quantum sensors that can measure the magnetic fields generated by muscle activity (magnetomyography, MMG) without skin contact. This contact-free alternative to surface electromyography (EMG) has remained mostly confined to magnetically shielded rooms and simple tasks, limiting translation. Here, we show that a portable OPM-MMG system operated inside a compact magnetic shield can robustly decode fine finger movements and recover EMG-like muscle activation patterns. Eight participants executed flexion- extension combinations spanning 15 finger actions while we recorded triaxial MMG and bipolar EMG concurrently. MMG supported robust multi-class and finger-specific decoding, recovering a representational geometry that closely matched EMG. Orientation analyses showed that components orthogonal to the muscle axis contributed most significantly to discriminability, providing actionable guidance for OPM array design. Our results show that OPM-MMG in a portable shielded environment preserves task-relevant neuromuscular information and approaches EMG-level structure without skin contact or a magnetically shielded room. Our findings open a path toward hygienic, rapid-setup assessment and human-machine interfacing in clinics and rehabilitation settings.