Oliver Kohl
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Oliver Kohl has a total of 45 co-authors. Here are the top co-authors by number of shared publications:
- Chetan Gohil
Research Department of Epilepsy, UCL Queen Square Institute of Neurology, University College London, United Kingdom
8 shared publications - Anna Christina Nobre
Oxford Centre for Human Brain Activity, Wellcome Centre for Integrative Neuroimaging, Department of Psychiatry, University of Oxford
6 shared publications - Mark W. Woolrich
Oxford Centre for Human Brain Activity, Wellcome Centre for Integrative Neuroimaging, Department of Psychiatry, University of Oxford
6 shared publications - Andrew J. Quinn
Centre for Human Brain Health, School of Psychology, University of Birmingham
5 shared publications - Jemma Pitt
Oxford Centre for Human Brain Activity, Wellcome Centre for Integrative Neuroimaging, Department of Psychiatry, University of Oxford
4 shared publications - Mats Willem Jacqueline van Es
University of Oxford
4 shared publications
Latest Publications
- Open Access
- 12 October 2026
- Published by eLife Sciences Publications, Ltd in eLife
- Vol. 14
Robust and replicable effects of ageing on resting-state brain electrophysiology measured with MEG
Non-invasive recordings of brain electrophysiology provide credible insights into decline in neuronal functioning with age. New approaches are required to translate these results into compelling clinical metrics and meet the global challenge of preserving brain health in ageing. Changes in neuronal dynamics with ageing are observable in the power spectra of EEG and MEG recordings. Highly promising candidates for electrophysiological markers have been identified, but progress is hindered by substantial methodological variability across studies. This makes it challenging to establish a clear consensus on the frequency, location, and direction of any single reported effect within the larger body of research. We estimate a full-frequency whole-head profile of the ageing effect on eyes-closed resting-state MEG using the GLM-Spectrum. This data representation is easily sharable, facilitates meta-analyses, and provides a framework for estimating statistical power of age effects for future study planning. We use this to show that the effect of age replicates across open-access MEG datasets and is robust to modelling of common covariates. Distinct components within the full-frequency profile have different effect sizes, indicating that sample-size planning for ageing effects must consider the specific features of interest. The frequency profile of ageing is strongly robust to a range of common covariates and partially robust to modelling of grey matter volume. We establish that what seems a well-powered study may become underpowered when analyses target an age effect that is linearly separable from an age-relevant covariate such as grey matter volume. These results provide a pathway towards formal comparison and assessment of candidate markers for brain health in ageing.
- 12 October 2026
- Published by eLife Sciences Publications, Ltd in eLife
- Vol. 14
Open AccessRobust and replicable effects of ageing on resting-state brain electrophysiology measured with MEG
Non-invasive recordings of brain electrophysiology provide credible insights into decline in neuronal functioning with age. New approaches are required to translate these results into compelling clinical metrics and meet the global challenge of preserving brain health in ageing. Changes in neuronal dynamics with ageing are observable in the power spectra of EEG and MEG recordings. Highly promising candidates for electrophysiological markers have been identified, but progress is hindered by substantial methodological variability across studies. This makes it challenging to establish a clear consensus on the frequency, location, and direction of any single reported effect within the larger body of research. We estimate a full-frequency whole-head profile of the ageing effect on eyes-closed resting-state MEG using the GLM-Spectrum. This data representation is easily sharable, facilitates meta-analyses, and provides a framework for estimating statistical power of age effects for future study planning. We use this to show that the effect of age replicates across open-access MEG datasets and is robust to modelling of common covariates. Distinct components within the full-frequency profile have different effect sizes, indicating that sample-size planning for ageing effects must consider the specific features of interest. The frequency profile of ageing is strongly robust to a range of common covariates and partially robust to modelling of grey matter volume. We establish that what seems a well-powered study may become underpowered when analyses target an age effect that is linearly separable from an age-relevant covariate such as grey matter volume. These results provide a pathway towards formal comparison and assessment of candidate markers for brain health in ageing.
- Open Access
Role of GABA and NMDA receptors in shaping cortical timescales and large-scale network dynamics
Cortical brain regions integrate information across different timescales, ranging from fast sensory processing to longer integration windows, allowing cognitive functions like working memory. At the large-scale, brain regions organize into transient network states that rapidly switch over time and similarly contribute to cognition. Both cortical timescales and large-scale network dynamics are proposed to be determined by the balance between recurrent synaptic excitation and GABAergic inhibition. Here, we pharmacologically manipulated synaptic transmission at GABA A and NMDA receptors in 60 healthy male participants and acquired resting-state magnetoencephalography. Neuronal timescales followed a hierarchical gradient with shorter timescales in early sensory regions. Increasing GABAergic activity prolonged neuronal timescales across cortical regions. This effect was most prominent in the frontal default mode and in the dorsal attention network. Notably, dynamic network analyses revealed that the occurrence probability of the frontal default mode network increased, whereas the occurrence of the dorsal attention network was reduced. NMDA receptor modulation resulted in no significant changes. Together, these findings provide causal evidence that GABAergic inhibition is a key regulator of cortical temporal organization, linking microscale synaptic mechanisms to neuronal timescales and network dynamics that support diverse cognitive function.
Open AccessRole of GABA and NMDA receptors in shaping cortical timescales and large-scale network dynamics
Cortical brain regions integrate information across different timescales, ranging from fast sensory processing to longer integration windows, allowing cognitive functions like working memory. At the large-scale, brain regions organize into transient network states that rapidly switch over time and similarly contribute to cognition. Both cortical timescales and large-scale network dynamics are proposed to be determined by the balance between recurrent synaptic excitation and GABAergic inhibition. Here, we pharmacologically manipulated synaptic transmission at GABA A and NMDA receptors in 60 healthy male participants and acquired resting-state magnetoencephalography. Neuronal timescales followed a hierarchical gradient with shorter timescales in early sensory regions. Increasing GABAergic activity prolonged neuronal timescales across cortical regions. This effect was most prominent in the frontal default mode and in the dorsal attention network. Notably, dynamic network analyses revealed that the occurrence probability of the frontal default mode network increased, whereas the occurrence of the dorsal attention network was reduced. NMDA receptor modulation resulted in no significant changes. Together, these findings provide causal evidence that GABAergic inhibition is a key regulator of cortical temporal organization, linking microscale synaptic mechanisms to neuronal timescales and network dynamics that support diverse cognitive function.
- Open Access
- 1 September 2025
- Published by Oxford University Press (OUP) in Brain Communications
- Vol. 7 (5)
Reduced beta bursting underpins loss of corticomuscular coherence in amyotrophic lateral sclerosis
Biomarkers of disease activity that holistically capture motor system dysfunction are needed to accelerate drug discovery in amyotrophic lateral sclerosis. Magnetoencephalography is a sensitive, non-invasive measure of cortical neurophysiology. Corticomuscular coherence reflects the functional coupling of cortical oscillations with downstream muscle activity recorded by electromyography. Cortical beta frequency bursting is known to represent a core feature of the neurophysiology underpinning movement. This study aimed to characterize disruption of beta frequency activity in both cortex and muscle to refine the understanding of corticomuscular coherence loss in amyotrophic lateral sclerosis. The study analysed 42 people living with amyotrophic lateral sclerosis and 33 healthy age-matched controls. Participants undertook an isometric hand gripping task during magnetoencephalography. Muscle contraction was measured using bipolar surface electromyography recordings at both forearms. All participants performed 120 trials of the gripper task bilaterally, and 60 trials unilaterally on each side. For each trial type, the mean corticomuscular coherence over trials was calculated for each participant and the groups were compared via cluster-based permutations tests. Beta burst metrics were calculated for the motor cortex (magnetoencephalography) and flexor forearm muscles (surface electromyography) including burst fractional occupancy, burst duration and amplitude. During muscular contraction, beta frequency corticomuscular coherence from the motor cortices contralateral to the gripper task was markedly reduced in amyotrophic lateral sclerosis patients, despite no significant difference in grip strength compared with controls. Source localization analysis showed globally reduced corticomuscular coherence in amyotrophic lateral sclerosis with significant differences in the motor regions contralateral to the engaged hand. There were no significant beta frequency activity changes in the engaged-hand electromyography signal in amyotrophic lateral sclerosis compared with controls. In contrast, analysis of the cortical motor regions revealed reduced rate of beta bursting and higher amplitude during the contraction phase of the task in amyotrophic lateral sclerosis. The corticomuscular coherence disruption in amyotrophic lateral sclerosis appears driven more by cerebral pathology than by muscle denervation. Equal grip strength during the task implies compensatory pathways in disease that are not captured by corticomuscular coherence. Interneuronal dysfunction may underlie the disruption to motor cortex beta bursting. Motor cortical beta frequency metrics have potential as secondary outcome measures in therapeutic trials and need exploration as prodromal markers in asymptomatic individuals genetically predisposed to amyotrophic lateral sclerosis.
- 1 September 2025
- Published by Oxford University Press (OUP) in Brain Communications
- Vol. 7 (5)
Open AccessReduced beta bursting underpins loss of corticomuscular coherence in amyotrophic lateral sclerosis
Biomarkers of disease activity that holistically capture motor system dysfunction are needed to accelerate drug discovery in amyotrophic lateral sclerosis. Magnetoencephalography is a sensitive, non-invasive measure of cortical neurophysiology. Corticomuscular coherence reflects the functional coupling of cortical oscillations with downstream muscle activity recorded by electromyography. Cortical beta frequency bursting is known to represent a core feature of the neurophysiology underpinning movement. This study aimed to characterize disruption of beta frequency activity in both cortex and muscle to refine the understanding of corticomuscular coherence loss in amyotrophic lateral sclerosis. The study analysed 42 people living with amyotrophic lateral sclerosis and 33 healthy age-matched controls. Participants undertook an isometric hand gripping task during magnetoencephalography. Muscle contraction was measured using bipolar surface electromyography recordings at both forearms. All participants performed 120 trials of the gripper task bilaterally, and 60 trials unilaterally on each side. For each trial type, the mean corticomuscular coherence over trials was calculated for each participant and the groups were compared via cluster-based permutations tests. Beta burst metrics were calculated for the motor cortex (magnetoencephalography) and flexor forearm muscles (surface electromyography) including burst fractional occupancy, burst duration and amplitude. During muscular contraction, beta frequency corticomuscular coherence from the motor cortices contralateral to the gripper task was markedly reduced in amyotrophic lateral sclerosis patients, despite no significant difference in grip strength compared with controls. Source localization analysis showed globally reduced corticomuscular coherence in amyotrophic lateral sclerosis with significant differences in the motor regions contralateral to the engaged hand. There were no significant beta frequency activity changes in the engaged-hand electromyography signal in amyotrophic lateral sclerosis compared with controls. In contrast, analysis of the cortical motor regions revealed reduced rate of beta bursting and higher amplitude during the contraction phase of the task in amyotrophic lateral sclerosis. The corticomuscular coherence disruption in amyotrophic lateral sclerosis appears driven more by cerebral pathology than by muscle denervation. Equal grip strength during the task implies compensatory pathways in disease that are not captured by corticomuscular coherence. Interneuronal dysfunction may underlie the disruption to motor cortex beta bursting. Motor cortical beta frequency metrics have potential as secondary outcome measures in therapeutic trials and need exploration as prodromal markers in asymptomatic individuals genetically predisposed to amyotrophic lateral sclerosis.
- Open Access
Robust and replicable effects of ageing on resting state brain electrophysiology measured with MEG
Non-invasive recordings using brain electrophysiology provide credible insights into decline in neuronal functioning with age. New approaches are required to translate these results into compelling clinical metrics and meet the global challenge of preserving brain health in ageing. Changes in neuronal dynamics with ageing are observable in the power spectra of EEG and MEG recordings. Highly promising candidates for electrophysiological markers have been identified, but progress is hindered by substantial methodological variability across studies. This makes it challenging to establish a clear consensus on the frequency, location, and direction of any single reported effect within the larger body of research. We estimate a full-frequency whole-head profile of the ageing effect on eyes-closed resting-state MEG using the GLM-Spectrum. This data representation is easily sharable, facilitates meta-analyses and provides a framework for estimating statistical power of age effects for future study planning. We use this to show that the effect of age replicates across open-access MEG datasets and is robust to modelling of common covariates. Distinct components within the full frequency profile have different effect sizes, indicating that sample-size planning for ageing effects must consider the specific features of interest. The frequency profile of ageing is strongly robust to a range of common covariates and partially robust to modelling of grey matter volume. We establish that what seems a well-powered study may become underpowered when analyses target an age effect that is linearly separable from an age-relevant covariate such as grey matter volume. These results provide a pathway towards formal comparison and assessment of candidate markers for brain health in ageing. Graphical Highlights: There is a strong effect of ageing across the adult lifespan on neuronal oscillations across a range of frequencies and spatial locations. The effect is replicable across multiple open-access datasets. The effect is robust to a range of sources of inter-subject variability. The age effect has heterogeneous effect size across space and frequency with implications for future study planning.
Open AccessRobust and replicable effects of ageing on resting state brain electrophysiology measured with MEG
Non-invasive recordings using brain electrophysiology provide credible insights into decline in neuronal functioning with age. New approaches are required to translate these results into compelling clinical metrics and meet the global challenge of preserving brain health in ageing. Changes in neuronal dynamics with ageing are observable in the power spectra of EEG and MEG recordings. Highly promising candidates for electrophysiological markers have been identified, but progress is hindered by substantial methodological variability across studies. This makes it challenging to establish a clear consensus on the frequency, location, and direction of any single reported effect within the larger body of research. We estimate a full-frequency whole-head profile of the ageing effect on eyes-closed resting-state MEG using the GLM-Spectrum. This data representation is easily sharable, facilitates meta-analyses and provides a framework for estimating statistical power of age effects for future study planning. We use this to show that the effect of age replicates across open-access MEG datasets and is robust to modelling of common covariates. Distinct components within the full frequency profile have different effect sizes, indicating that sample-size planning for ageing effects must consider the specific features of interest. The frequency profile of ageing is strongly robust to a range of common covariates and partially robust to modelling of grey matter volume. We establish that what seems a well-powered study may become underpowered when analyses target an age effect that is linearly separable from an age-relevant covariate such as grey matter volume. These results provide a pathway towards formal comparison and assessment of candidate markers for brain health in ageing. Graphical Highlights: There is a strong effect of ageing across the adult lifespan on neuronal oscillations across a range of frequencies and spatial locations. The effect is replicable across multiple open-access datasets. The effect is robust to a range of sources of inter-subject variability. The age effect has heterogeneous effect size across space and frequency with implications for future study planning.
- Open Access
- 3 July 2025
- Published by Oxford University Press (OUP) in Brain Communications
- Vol. 7 (4) , fcaf282
Changes in sensorimotor network dynamics in resting-state recordings in Parkinson’s disease
Non-invasive recordings of magnetoencephalography have been used for developing biomarkers for neural changes associated with Parkinson’s disease that can be measured across the entire course of the disease. These studies, however, have yielded inconsistent findings. Here, we investigated whether analysing motor cortical activity within the context of large-scale brain network activity provides a more sensitive marker of changes in Parkinson’s disease using magnetoencephalography. We extracted motor cortical beta power and beta bursts from resting-state magnetoencephalography scans of patients with Parkinson’s disease (N = 28) and well-matched healthy controls (N = 36). To situate beta bursts in their brain network contexts, we used a time-delay-embedded hidden Markov model to extract brain network activity and investigated co-occurrence patterns between brain networks and beta bursts. Parkinson’s disease was associated with decreased beta power in motor cortical power spectra, but no significant differences in motor cortical beta-burst dynamics occurred when using a conventional beta-burst analysis. Dynamics of a large-scale sensorimotor network extracted with the time-delay-embedded hidden Markov model approach revealed significant decreases in the occurrence of this network with Parkinson’s disease. By comparing conventional burst and time-delay-embedded hidden Markov model state occurrences, we observed that motor beta bursts occurred during both sensorimotor and non-sensorimotor network activations. When using the large-scale network information provided by the time-delay-embedded hidden Markov model to focus on bursts that were active during sensorimotor network activations, significant decreases in burst dynamics could be observed in patients with Parkinson’s disease. In conclusion, our findings suggest that decreased motor cortical beta power in Parkinson’s disease is prominently associated with changes in sensorimotor network dynamics using magnetoencephalography. Thus, investigating large-scale networks or considering the large-scale network context of motor cortical activations may be crucial for identifying alterations in the sensorimotor network that are prevalent in Parkinson’s disease and might help resolve contradicting findings in the literature.
- 3 July 2025
- Published by Oxford University Press (OUP) in Brain Communications
- Vol. 7 (4) , fcaf282
Open AccessChanges in sensorimotor network dynamics in resting-state recordings in Parkinson’s disease
Non-invasive recordings of magnetoencephalography have been used for developing biomarkers for neural changes associated with Parkinson’s disease that can be measured across the entire course of the disease. These studies, however, have yielded inconsistent findings. Here, we investigated whether analysing motor cortical activity within the context of large-scale brain network activity provides a more sensitive marker of changes in Parkinson’s disease using magnetoencephalography. We extracted motor cortical beta power and beta bursts from resting-state magnetoencephalography scans of patients with Parkinson’s disease (N = 28) and well-matched healthy controls (N = 36). To situate beta bursts in their brain network contexts, we used a time-delay-embedded hidden Markov model to extract brain network activity and investigated co-occurrence patterns between brain networks and beta bursts. Parkinson’s disease was associated with decreased beta power in motor cortical power spectra, but no significant differences in motor cortical beta-burst dynamics occurred when using a conventional beta-burst analysis. Dynamics of a large-scale sensorimotor network extracted with the time-delay-embedded hidden Markov model approach revealed significant decreases in the occurrence of this network with Parkinson’s disease. By comparing conventional burst and time-delay-embedded hidden Markov model state occurrences, we observed that motor beta bursts occurred during both sensorimotor and non-sensorimotor network activations. When using the large-scale network information provided by the time-delay-embedded hidden Markov model to focus on bursts that were active during sensorimotor network activations, significant decreases in burst dynamics could be observed in patients with Parkinson’s disease. In conclusion, our findings suggest that decreased motor cortical beta power in Parkinson’s disease is prominently associated with changes in sensorimotor network dynamics using magnetoencephalography. Thus, investigating large-scale networks or considering the large-scale network context of motor cortical activations may be crucial for identifying alterations in the sensorimotor network that are prevalent in Parkinson’s disease and might help resolve contradicting findings in the literature.
- Open Access
Effects of age on resting-state cortical networks
Understanding how ageing affects brain function remains a central challenge in neuroscience. Electrophysiological brain imaging techniques provide a near-direct measure of neuronal activity, which is useful for characterising neurophysiological health. They offer us the ability to track large-scale networks of functional activity with high temporal precision. The effects of healthy ageing on these networks remain poorly understood, in part due to small sample sizes and limited control for confounding factors in previous studies. Here, we analysed resting-state source-reconstructed magnetoencephalography (MEG) data from a large cross-sectional cohort of healthy adults (N=612, 18-88 years old) to characterise the effect of age using not only time-averaged (static), but also transient (dynamic) network activity. We examined time-averaged power and coherence across canonical frequency bands (δ, θ, α, β, γ), as well as transient network dynamics identified using Hidden Markov Modelling. We included many confounding variables known to be affected by age, such as brain volume, as well as head size and position, which have previously been overlooked. Ageing was associated with frequency-specific changes in oscillatory power, with decreases in low-frequency (δ, θ) power and increases in high-frequency (β) power. Coherence increased across all frequency bands and was positively associated with cognitive performance. Transient network analyses additionally revealed that frontal network occurrences declined with age, with evidence suggesting a compensatory role in supporting cognition. These findings provide a more comprehensive electrophysiological signature for healthy ageing and establish a baseline for detecting pathological change.
Open AccessEffects of age on resting-state cortical networks
Understanding how ageing affects brain function remains a central challenge in neuroscience. Electrophysiological brain imaging techniques provide a near-direct measure of neuronal activity, which is useful for characterising neurophysiological health. They offer us the ability to track large-scale networks of functional activity with high temporal precision. The effects of healthy ageing on these networks remain poorly understood, in part due to small sample sizes and limited control for confounding factors in previous studies. Here, we analysed resting-state source-reconstructed magnetoencephalography (MEG) data from a large cross-sectional cohort of healthy adults (N=612, 18-88 years old) to characterise the effect of age using not only time-averaged (static), but also transient (dynamic) network activity. We examined time-averaged power and coherence across canonical frequency bands (δ, θ, α, β, γ), as well as transient network dynamics identified using Hidden Markov Modelling. We included many confounding variables known to be affected by age, such as brain volume, as well as head size and position, which have previously been overlooked. Ageing was associated with frequency-specific changes in oscillatory power, with decreases in low-frequency (δ, θ) power and increases in high-frequency (β) power. Coherence increased across all frequency bands and was positively associated with cognitive performance. Transient network analyses additionally revealed that frontal network occurrences declined with age, with evidence suggesting a compensatory role in supporting cognition. These findings provide a more comprehensive electrophysiological signature for healthy ageing and establish a baseline for detecting pathological change.
- Open Access
- 3 May 2024
- Published by Oxford University Press (OUP) in Brain Communications
- Vol. 6 (3) , fcae164
The cortical neurophysiological signature of amyotrophic lateral sclerosis
The progressive loss of motor function characteristic of amyotrophic lateral sclerosis is associated with widespread cortical pathology extending beyond primary motor regions. Increasing muscle weakness reflects a dynamic, variably compensated brain network disorder. In the quest for biomarkers to accelerate therapeutic assessment, the high temporal resolution of magnetoencephalography is uniquely able to non-invasively capture micro-magnetic fields generated by neuronal activity across the entire cortex simultaneously. This study examined task-free magnetoencephalography to characterize the cortical oscillatory signature of amyotrophic lateral sclerosis for having potential as a pharmacodynamic biomarker. Eight to ten minutes of magnetoencephalography in the task-free, eyes-open state was recorded in amyotrophic lateral sclerosis (n = 36) and healthy age-matched controls (n = 51), followed by a structural MRI scan for co-registration. Extracted magnetoencephalography metrics from the delta, theta, alpha, beta, low-gamma, high-gamma frequency bands included oscillatory power (regional activity), 1/f exponent (complexity) and amplitude envelope correlation (connectivity). Groups were compared using a permutation-based general linear model with correction for multiple comparisons and confounders. To test whether the extracted metrics could predict disease severity, a random forest regression model was trained and evaluated using nested leave-one-out cross-validation. Amyotrophic lateral sclerosis was characterized by reduced sensorimotor beta band and increased high-gamma band power. Within the premotor cortex, increased disability was associated with a reduced 1/f exponent. Increased disability was more widely associated with increased global connectivity in the delta, theta and high-gamma bands. Intra-hemispherically, increased disability scores were particularly associated with increases in temporal connectivity and inter-hemispherically with increases in frontal and occipital connectivity. The random forest model achieved a coefficient of determination (R2) of 0.24. The combined reduction in cortical sensorimotor beta and rise in gamma power is compatible with the established hypothesis of loss of inhibitory, GABAergic interneuronal circuits in pathogenesis. A lower 1/f exponent potentially reflects a more excitable cortex and a pathology unique to amyotrophic lateral sclerosis when considered with the findings published in other neurodegenerative disorders. Power and complexity changes corroborate with the results from paired-pulse transcranial magnetic stimulation. Increased magnetoencephalography connectivity in worsening disability is thought to represent compensatory responses to a failing motor system. Restoration of cortical beta and gamma band power has significant potential to be tested in an experimental medicine setting. Magnetoencephalography-based measures have potential as sensitive outcome measures of therapeutic benefit in drug trials and may have a wider diagnostic value with further study, including as predictive markers in asymptomatic carriers of disease-causing genetic variants.
- 3 May 2024
- Published by Oxford University Press (OUP) in Brain Communications
- Vol. 6 (3) , fcae164
Open AccessThe cortical neurophysiological signature of amyotrophic lateral sclerosis
The progressive loss of motor function characteristic of amyotrophic lateral sclerosis is associated with widespread cortical pathology extending beyond primary motor regions. Increasing muscle weakness reflects a dynamic, variably compensated brain network disorder. In the quest for biomarkers to accelerate therapeutic assessment, the high temporal resolution of magnetoencephalography is uniquely able to non-invasively capture micro-magnetic fields generated by neuronal activity across the entire cortex simultaneously. This study examined task-free magnetoencephalography to characterize the cortical oscillatory signature of amyotrophic lateral sclerosis for having potential as a pharmacodynamic biomarker. Eight to ten minutes of magnetoencephalography in the task-free, eyes-open state was recorded in amyotrophic lateral sclerosis (n = 36) and healthy age-matched controls (n = 51), followed by a structural MRI scan for co-registration. Extracted magnetoencephalography metrics from the delta, theta, alpha, beta, low-gamma, high-gamma frequency bands included oscillatory power (regional activity), 1/f exponent (complexity) and amplitude envelope correlation (connectivity). Groups were compared using a permutation-based general linear model with correction for multiple comparisons and confounders. To test whether the extracted metrics could predict disease severity, a random forest regression model was trained and evaluated using nested leave-one-out cross-validation. Amyotrophic lateral sclerosis was characterized by reduced sensorimotor beta band and increased high-gamma band power. Within the premotor cortex, increased disability was associated with a reduced 1/f exponent. Increased disability was more widely associated with increased global connectivity in the delta, theta and high-gamma bands. Intra-hemispherically, increased disability scores were particularly associated with increases in temporal connectivity and inter-hemispherically with increases in frontal and occipital connectivity. The random forest model achieved a coefficient of determination (R2) of 0.24. The combined reduction in cortical sensorimotor beta and rise in gamma power is compatible with the established hypothesis of loss of inhibitory, GABAergic interneuronal circuits in pathogenesis. A lower 1/f exponent potentially reflects a more excitable cortex and a pathology unique to amyotrophic lateral sclerosis when considered with the findings published in other neurodegenerative disorders. Power and complexity changes corroborate with the results from paired-pulse transcranial magnetic stimulation. Increased magnetoencephalography connectivity in worsening disability is thought to represent compensatory responses to a failing motor system. Restoration of cortical beta and gamma band power has significant potential to be tested in an experimental medicine setting. Magnetoencephalography-based measures have potential as sensitive outcome measures of therapeutic benefit in drug trials and may have a wider diagnostic value with further study, including as predictive markers in asymptomatic carriers of disease-causing genetic variants.
- Open Access
Changes in sensorimotor network dynamics in resting-state recordings in Parkinson’s Disease
Non-invasive recordings of magnetoencephalography (MEG) have been used for developing biomarkers for neural changes associated with Parkinson’s Disease (PD) but have yielded inconsistent findings. Here, we investigated whether analysing motor cortical activity within the context of large-scale brain network activity provides a more sensitive marker of changes in PD using MEG. We extracted motor cortical beta power and beta bursts from resting-state MEG scans of individuals with PD (N=28) and well-matched healthy controls (N=36). To situate beta bursts in their brain network contexts, we used a time-delay embedded Hidden Markov Model (TDE-HMM) to extract brain network activity and investigated co-occurrence patterns between brain networks and beta bursts. PD was associated with decreased beta power in motor cortex and decreased occurrences of the sensorimotor network, while motor cortical beta-burst dynamics were not changed. By comparing conventional burst and large-scale network occurrences, we observed that motor beta bursts occurred during both sensorimotor network and non-sensorimotor network activations. When using the large-scale network information provided by the TDE-HMM to focus on bursts that were active during sensorimotor network activations, significant decreases in burst dynamics could be observed in individuals with PD. In conclusion, our findings suggest that decreased motor cortical beta power in PD is prominently associated with changes in sensorimotor network dynamics using MEG. Thus, investigating large-scale networks or considering the large-scale network context of motor cortical activations may be crucial for identifying alterations in the sensorimotor network that are prevalent in PD, and might help resolve contradicting findings in the literature. Highlights Sensorimotor network occurrences are decreased in Parkinson’s Disease. Motor cortical beta bursts occur during both sensorimotor network and non-sensorimotor network activations. Focusing on motor beta bursts occurring during sensorimotor network activations enables for better discrimination between controls and individuals with PD. The spatiotemporal details provided by large-scale network analysis may help to overcome discrepancies found in the PD literature.
Open AccessChanges in sensorimotor network dynamics in resting-state recordings in Parkinson’s Disease
Non-invasive recordings of magnetoencephalography (MEG) have been used for developing biomarkers for neural changes associated with Parkinson’s Disease (PD) but have yielded inconsistent findings. Here, we investigated whether analysing motor cortical activity within the context of large-scale brain network activity provides a more sensitive marker of changes in PD using MEG. We extracted motor cortical beta power and beta bursts from resting-state MEG scans of individuals with PD (N=28) and well-matched healthy controls (N=36). To situate beta bursts in their brain network contexts, we used a time-delay embedded Hidden Markov Model (TDE-HMM) to extract brain network activity and investigated co-occurrence patterns between brain networks and beta bursts. PD was associated with decreased beta power in motor cortex and decreased occurrences of the sensorimotor network, while motor cortical beta-burst dynamics were not changed. By comparing conventional burst and large-scale network occurrences, we observed that motor beta bursts occurred during both sensorimotor network and non-sensorimotor network activations. When using the large-scale network information provided by the TDE-HMM to focus on bursts that were active during sensorimotor network activations, significant decreases in burst dynamics could be observed in individuals with PD. In conclusion, our findings suggest that decreased motor cortical beta power in PD is prominently associated with changes in sensorimotor network dynamics using MEG. Thus, investigating large-scale networks or considering the large-scale network context of motor cortical activations may be crucial for identifying alterations in the sensorimotor network that are prevalent in PD, and might help resolve contradicting findings in the literature. Highlights Sensorimotor network occurrences are decreased in Parkinson’s Disease. Motor cortical beta bursts occur during both sensorimotor network and non-sensorimotor network activations. Focusing on motor beta bursts occurring during sensorimotor network activations enables for better discrimination between controls and individuals with PD. The spatiotemporal details provided by large-scale network analysis may help to overcome discrepancies found in the PD literature.
- Open Access
Dynamic Network Analysis of Electrophysiological Task Data
An important approach for studying the human brain is to use functional neuroimaging combined with a task. In electrophysiological data this often involves a time-frequency analysis, in which recorded brain activity is time-frequency transformed and epoched around task events of interest, followed by trial-averaging of the power. Whilst this simple approach can reveal fast oscillatory dynamics, the brain regions are analysed one at a time. This causes difficulties for interpretation and a debilitating number of multiple comparisons. In addition, it is now recognised that the brain responds to tasks through the coordinated activity of networks of brain areas. As such, techniques that take a whole-brain network perspective are needed. Here, we show how the oscillatory task responses from conventional time-frequency approaches, can be represented more parsimoniously at the network level using two state-of-the-art methods: the HMM (Hidden Markov Model) and DyNeMo (Dynamic Network Modes). Both methods reveal frequency-resolved networks of oscillatory activity with millisecond resolution. Comparing DyNeMo, HMM and traditional oscillatory response analysis, we show DyNeMo can identify task activations/deactivations that the other approaches fail to detect. DyNeMo offers a powerful new method for analysing task data from the perspective of dynamic brain networks. Highlights We show how oscillatory task response analysis can be carried out at the network level using two state-of-the-art methods: the Hidden Markov Model (HMM) and Dynamics Network Modes (DyNeMo). The HMM and DyNeMo can identify oscillatory task responses that conventional time-frequency methods fail to detect. DyNeMo provides a more interpretable and precise network decomposition of the data compared to the HMM. The dataset and (Python) scripts used to perform dynamic network analysis on task data are made publicly available.
Open AccessDynamic Network Analysis of Electrophysiological Task Data
An important approach for studying the human brain is to use functional neuroimaging combined with a task. In electrophysiological data this often involves a time-frequency analysis, in which recorded brain activity is time-frequency transformed and epoched around task events of interest, followed by trial-averaging of the power. Whilst this simple approach can reveal fast oscillatory dynamics, the brain regions are analysed one at a time. This causes difficulties for interpretation and a debilitating number of multiple comparisons. In addition, it is now recognised that the brain responds to tasks through the coordinated activity of networks of brain areas. As such, techniques that take a whole-brain network perspective are needed. Here, we show how the oscillatory task responses from conventional time-frequency approaches, can be represented more parsimoniously at the network level using two state-of-the-art methods: the HMM (Hidden Markov Model) and DyNeMo (Dynamic Network Modes). Both methods reveal frequency-resolved networks of oscillatory activity with millisecond resolution. Comparing DyNeMo, HMM and traditional oscillatory response analysis, we show DyNeMo can identify task activations/deactivations that the other approaches fail to detect. DyNeMo offers a powerful new method for analysing task data from the perspective of dynamic brain networks. Highlights We show how oscillatory task response analysis can be carried out at the network level using two state-of-the-art methods: the Hidden Markov Model (HMM) and Dynamics Network Modes (DyNeMo). The HMM and DyNeMo can identify oscillatory task responses that conventional time-frequency methods fail to detect. DyNeMo provides a more interpretable and precise network decomposition of the data compared to the HMM. The dataset and (Python) scripts used to perform dynamic network analysis on task data are made publicly available.
- Open Access
The GLM-Spectrum: A multilevel framework for spectrum analysis with covariate and confound modelling
The frequency spectrum is a central method for representing the dynamics within electrophysiological data. Some widely used spectrum estimators make use of averaging across time segments to reduce noise in the final spectrum. The core of this approach has not changed substantially since the 1960s, though many advances in the field of regression modelling and statistics have been made during this time. Here, we propose a new approach, the General Linear Model (GLM) Spectrum, which reframes time averaged spectral estimation as multiple regression. This brings several benefits, including the ability to do confound modelling, hierarchical modelling and significance testing via non-parametric statistics. We apply the approach to a dataset of EEG recordings of participants who alternate between eyes-open and eyes-closed resting state. The GLM-Spectrum can model both conditions, quantify their differences, and perform denoising through confound regression in a single step. This application is scaled up from a single channel to a whole head recording and, finally, applied to quantify age differences across a large group-level dataset. We show that the GLM-Spectrum lends itself to rigorous modelling of within- and between-subject contrasts as well as their interactions, and that the use of model-projected spectra provides an intuitive visualisation. The GLM-Spectrum is a flexible framework for robust multi-level analysis of power spectra, with adaptive covariance and confound modelling.
Open AccessThe GLM-Spectrum: A multilevel framework for spectrum analysis with covariate and confound modelling
The frequency spectrum is a central method for representing the dynamics within electrophysiological data. Some widely used spectrum estimators make use of averaging across time segments to reduce noise in the final spectrum. The core of this approach has not changed substantially since the 1960s, though many advances in the field of regression modelling and statistics have been made during this time. Here, we propose a new approach, the General Linear Model (GLM) Spectrum, which reframes time averaged spectral estimation as multiple regression. This brings several benefits, including the ability to do confound modelling, hierarchical modelling and significance testing via non-parametric statistics. We apply the approach to a dataset of EEG recordings of participants who alternate between eyes-open and eyes-closed resting state. The GLM-Spectrum can model both conditions, quantify their differences, and perform denoising through confound regression in a single step. This application is scaled up from a single channel to a whole head recording and, finally, applied to quantify age differences across a large group-level dataset. We show that the GLM-Spectrum lends itself to rigorous modelling of within- and between-subject contrasts as well as their interactions, and that the use of model-projected spectra provides an intuitive visualisation. The GLM-Spectrum is a flexible framework for robust multi-level analysis of power spectra, with adaptive covariance and confound modelling.