Paul J. N. Brodersen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Paul J. N. Brodersen has a total of 35 co-authors. Here are the top co-authors by number of shared publications:
- Colin J. Akerman
Department of Pharmacology, University of Oxford, Oxford, UK
5 shared publications - Christopher W. Thomas
COMPASS Pathways plc, London, UK
2 shared publications Hannah Alfonsa
Department of Pharmacology, University of Oxford, Mansfield Road, Oxford, OX1 3QT, UK
2 shared publicationsLewis Taylor
Nuffield Department of Clinical Neurosciences, University of Oxford; John Radcliffe Hospital, Oxford, United Kingdom
2 shared publications- Lukas Bernhard Krone
Department of Physiology, Anatomy and Genetics, University of Oxford
2 shared publications - Mathilde C. C. Guillaumin
Brain Network Dynamics Unit, Nuffield Department for Clinical Neuroscience, University of Oxford, Oxford, UK
2 shared publications
Latest Publications
- Open Access
Opposing intrinsic and synaptic plasticity mechanisms stabilize altered cortical networks during sleep deprivation
Sleep deprivation alters brain activity and impairs performance, yet many aspects of behaviour are preserved in this state. In the cortex for example, sleep deprivation increases neuronal firing rates and low-frequency oscillatory activity, but cortical circuits continue to process information. Recent work has implicated synaptic inhibition in these changes, with sleep deprivation causing cortical GABA A receptor (GABA A R) signalling to become depolarizing due to changes in chloride gradients that determine the GABA A R reversal potential (E GABAAR ). The impact in the intact brain remains unclear, however, as both the degree and effects of E GABAAR changes depend on network activity and intrinsic properties of neurons. To address this, we perform in vivo gramicidin recordings from cortical pyramidal neurons in sleep-deprived mice. We reveal that synaptic E GABAAR is depolarised to values sufficient to support bona fide excitatory GABA A R signalling, through combined network-dependent and cell-autonomous effects. In the face of this depolarizing drive, cortical neurons engage spike threshold adaptation mechanisms that limit their excitability. These opposing effects stabilise activity in simulations of sleep-deprived cortical networks and reproduce the increased firing rates and low-frequency oscillatory activity. This interplay between intrinsic and synaptic mechanisms may help maintain network stability during sleep deprivation, while reducing flexibility for subsequent adaptation.
Open AccessOpposing intrinsic and synaptic plasticity mechanisms stabilize altered cortical networks during sleep deprivation
Sleep deprivation alters brain activity and impairs performance, yet many aspects of behaviour are preserved in this state. In the cortex for example, sleep deprivation increases neuronal firing rates and low-frequency oscillatory activity, but cortical circuits continue to process information. Recent work has implicated synaptic inhibition in these changes, with sleep deprivation causing cortical GABA A receptor (GABA A R) signalling to become depolarizing due to changes in chloride gradients that determine the GABA A R reversal potential (E GABAAR ). The impact in the intact brain remains unclear, however, as both the degree and effects of E GABAAR changes depend on network activity and intrinsic properties of neurons. To address this, we perform in vivo gramicidin recordings from cortical pyramidal neurons in sleep-deprived mice. We reveal that synaptic E GABAAR is depolarised to values sufficient to support bona fide excitatory GABA A R signalling, through combined network-dependent and cell-autonomous effects. In the face of this depolarizing drive, cortical neurons engage spike threshold adaptation mechanisms that limit their excitability. These opposing effects stabilise activity in simulations of sleep-deprived cortical networks and reproduce the increased firing rates and low-frequency oscillatory activity. This interplay between intrinsic and synaptic mechanisms may help maintain network stability during sleep deprivation, while reducing flexibility for subsequent adaptation.
- Open Access
Automated sleep scoring in hibernating and non-hibernating American black bears
Hibernating bears show remarkable metabolic suppression. Their decline in core body temperature (T b ) is moderate (from 38°C to 30-35°C), but their metabolism declines as much as 75%. To understand the role of sleep in this hypometabolic state, we recorded biotelemetrically EEG, EOG and EMG data over 3500 days from 16 captive American black bears in and out of hibernation under semi-natural conditions. This data set is too large to score manually for Wake, REM- and NREM sleep, so we tested two machine learning classifiers: (1) Somnotate trained on multiple one-day recordings, and (2) Somnivore, trained on a small subset from each recording. As automated scoring methods have not been applied to hibernating species before, a major concern is the effect changing brain temperature has on the EEG and on the machine learning based detection. Therefore, we selected reference data using consensus by 3 manual sleep scorers from each of 6 bears, two one-day recordings at the highest and lowest body temperatures during hibernation when T b was oscillating in multiday cycles, and a non-hibernating one-day recording in summer. Somnotate results were excellent when trained separately for hibernating and non-hibernating data. Training Somnotate separately for high and low T b within hibernation did not improve results further. Sleep times in hibernation were about 2x that in summer for both automated scores and manual scores (p0.05), but a small overestimate of sleep time in summer (p<0.05). Both applications yielded F-measures against manual scores in the 0.90-0.98 range. Outliers in the 0.67-0.88 range were correlated between the two applications, indicating that specific files are more challenging to annotate. We conclude that both applications have accuracies approaching that of manual scorers when trained on high quality data.
Open AccessAutomated sleep scoring in hibernating and non-hibernating American black bears
Hibernating bears show remarkable metabolic suppression. Their decline in core body temperature (T b ) is moderate (from 38°C to 30-35°C), but their metabolism declines as much as 75%. To understand the role of sleep in this hypometabolic state, we recorded biotelemetrically EEG, EOG and EMG data over 3500 days from 16 captive American black bears in and out of hibernation under semi-natural conditions. This data set is too large to score manually for Wake, REM- and NREM sleep, so we tested two machine learning classifiers: (1) Somnotate trained on multiple one-day recordings, and (2) Somnivore, trained on a small subset from each recording. As automated scoring methods have not been applied to hibernating species before, a major concern is the effect changing brain temperature has on the EEG and on the machine learning based detection. Therefore, we selected reference data using consensus by 3 manual sleep scorers from each of 6 bears, two one-day recordings at the highest and lowest body temperatures during hibernation when T b was oscillating in multiday cycles, and a non-hibernating one-day recording in summer. Somnotate results were excellent when trained separately for hibernating and non-hibernating data. Training Somnotate separately for high and low T b within hibernation did not improve results further. Sleep times in hibernation were about 2x that in summer for both automated scores and manual scores (p0.05), but a small overestimate of sleep time in summer (p<0.05). Both applications yielded F-measures against manual scores in the 0.90-0.98 range. Outliers in the 0.67-0.88 range were correlated between the two applications, indicating that specific files are more challenging to annotate. We conclude that both applications have accuracies approaching that of manual scorers when trained on high quality data.
- Open Access
- 17 January 2024
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 20 (1) , e1011793
Somnotate: A probabilistic sleep stage classifier for studying vigilance state transitions
Electrophysiological recordings from freely behaving animals are a widespread and powerful mode of investigation in sleep research. These recordings generate large amounts of data that require sleep stage annotation (polysomnography), in which the data is parcellated according to three vigilance states: awake, rapid eye movement (REM) sleep, and non-REM (NREM) sleep. Manual and current computational annotation methods ignore intermediate states because the classification features become ambiguous, even though intermediate states contain important information regarding vigilance state dynamics. To address this problem, we have developed "Somnotate"—a probabilistic classifier based on a combination of linear discriminant analysis (LDA) with a hidden Markov model (HMM). First we demonstrate that Somnotate sets new standards in polysomnography, exhibiting annotation accuracies that exceed human experts on mouse electrophysiological data, remarkable robustness to errors in the training data, compatibility with different recording configurations, and an ability to maintain high accuracy during experimental interventions. However, the key feature of Somnotate is that it quantifies and reports the certainty of its annotations. We leverage this feature to reveal that many intermediate vigilance states cluster around state transitions, whereas others correspond to failed attempts to transition. This enables us to show for the first time that the success rates of different types of transition are differentially affected by experimental manipulations and can explain previously observed sleep patterns. Somnotate is open-source and has the potential to both facilitate the study of sleep stage transitions and offer new insights into the mechanisms underlying sleep-wake dynamics.
- 17 January 2024
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 20 (1) , e1011793
Open AccessSomnotate: A probabilistic sleep stage classifier for studying vigilance state transitions
Electrophysiological recordings from freely behaving animals are a widespread and powerful mode of investigation in sleep research. These recordings generate large amounts of data that require sleep stage annotation (polysomnography), in which the data is parcellated according to three vigilance states: awake, rapid eye movement (REM) sleep, and non-REM (NREM) sleep. Manual and current computational annotation methods ignore intermediate states because the classification features become ambiguous, even though intermediate states contain important information regarding vigilance state dynamics. To address this problem, we have developed "Somnotate"—a probabilistic classifier based on a combination of linear discriminant analysis (LDA) with a hidden Markov model (HMM). First we demonstrate that Somnotate sets new standards in polysomnography, exhibiting annotation accuracies that exceed human experts on mouse electrophysiological data, remarkable robustness to errors in the training data, compatibility with different recording configurations, and an ability to maintain high accuracy during experimental interventions. However, the key feature of Somnotate is that it quantifies and reports the certainty of its annotations. We leverage this feature to reveal that many intermediate vigilance states cluster around state transitions, whereas others correspond to failed attempts to transition. This enables us to show for the first time that the success rates of different types of transition are differentially affected by experimental manipulations and can explain previously observed sleep patterns. Somnotate is open-source and has the potential to both facilitate the study of sleep stage transitions and offer new insights into the mechanisms underlying sleep-wake dynamics.
- Open Access
- 19 July 2023
- Published by The Open Journal in Journal of Open Source Software
- Vol. 8 (87) , 5372
Netgraph: Publication-quality Network Visualisations in Python
The empirical study and scholarly analysis of networks have increased manifold in recent decades, fuelled by the new prominence of network structures in our lives (the web, social networks, artificial neural networks, ecological networks, etc.) and the data available on them. While there are several comprehensive Python libraries for network analysis such as NetworkX (Hagberg et al., 2008), igraph (Csardi & Nepusz, 2006), and graph-tool (Peixoto, 2014), their inbuilt visualisation capabilities lag behind specialised software solutions such as Graphviz (Ellson et al., 2002), Cytoscape (Shannon et al., 2003), and Gephi (Bastian et al., 2009). However, although Python bindings for these applications exist in the form of PyGraphviz, py4cytoscape, and GephiStreamer, respectively, their outputs are not manipulable Python objects, which restricts customisation, limits their extensibility, and prevents a seamless integration within a wider Python application. Summary Netgraph is a Python library that aims to complement the existing network analysis libraries with publication-quality visualisations within the Python ecosystem. To facilitate a seamless integration, Netgraph supports a variety of input formats, including NetworkX, igraph, and graph-tool Graph objects. At the time of writing, Netgraph provides the following node layout algorithms:
- 19 July 2023
- Published by The Open Journal in Journal of Open Source Software
- Vol. 8 (87) , 5372
Open AccessNetgraph: Publication-quality Network Visualisations in Python
The empirical study and scholarly analysis of networks have increased manifold in recent decades, fuelled by the new prominence of network structures in our lives (the web, social networks, artificial neural networks, ecological networks, etc.) and the data available on them. While there are several comprehensive Python libraries for network analysis such as NetworkX (Hagberg et al., 2008), igraph (Csardi & Nepusz, 2006), and graph-tool (Peixoto, 2014), their inbuilt visualisation capabilities lag behind specialised software solutions such as Graphviz (Ellson et al., 2002), Cytoscape (Shannon et al., 2003), and Gephi (Bastian et al., 2009). However, although Python bindings for these applications exist in the form of PyGraphviz, py4cytoscape, and GephiStreamer, respectively, their outputs are not manipulable Python objects, which restricts customisation, limits their extensibility, and prevents a seamless integration within a wider Python application. Summary Netgraph is a Python library that aims to complement the existing network analysis libraries with publication-quality visualisations within the Python ecosystem. To facilitate a seamless integration, Netgraph supports a variety of input formats, including NetworkX, igraph, and graph-tool Graph objects. At the time of writing, Netgraph provides the following node layout algorithms:
- Open Access
Active cortical networks promote shunting fast synaptic inhibition in vivo
Fast synaptic inhibition determines neuronal response properties in the mammalian brain and is mediated by chloride-permeable ionotropic GABA-A receptors (GABA A Rs). Despite their fundamental role, it is still not known how GABA A Rs signal in the intact brain. Here we use in vivo gramicidin recordings to investigate synaptic GABA A R signaling in mouse cortical pyramidal neurons under conditions that preserve native transmembrane chloride gradients. In anaesthetized cortex, synaptic GABA A Rs exert classic hyperpolarizing effects. In contrast, GABA A R-mediated synaptic signaling in awake cortex is found to be predominantly shunting. This is due to more depolarized GABA A R equilibrium potentials (E GABAAR ), which are shown to result from the high levels of synaptic activity that characterize awake cortical networks. The E GABAAR observed in awake cortex can facilitate the decoupling of local networks, which improves the ability of the network to discriminate stimuli. Our findings therefore suggest that GABA A R signaling adapts to optimize cortical functions.
Open AccessActive cortical networks promote shunting fast synaptic inhibition in vivo
Fast synaptic inhibition determines neuronal response properties in the mammalian brain and is mediated by chloride-permeable ionotropic GABA-A receptors (GABA A Rs). Despite their fundamental role, it is still not known how GABA A Rs signal in the intact brain. Here we use in vivo gramicidin recordings to investigate synaptic GABA A R signaling in mouse cortical pyramidal neurons under conditions that preserve native transmembrane chloride gradients. In anaesthetized cortex, synaptic GABA A Rs exert classic hyperpolarizing effects. In contrast, GABA A R-mediated synaptic signaling in awake cortex is found to be predominantly shunting. This is due to more depolarized GABA A R equilibrium potentials (E GABAAR ), which are shown to result from the high levels of synaptic activity that characterize awake cortical networks. The E GABAAR observed in awake cortex can facilitate the decoupling of local networks, which improves the ability of the network to discriminate stimuli. Our findings therefore suggest that GABA A R signaling adapts to optimize cortical functions.
- Open Access
Biologically plausible unsupervised learning in neural networks with sparse and asymmetric connectivity
In the search for biologically plausible but mathematically precise theories of learning in the brain, recent studies have begun to investigate how key assumptions underlying algorithms for supervised learning in artificial neural networks can be relaxed in biologically plausible ways. Turning to unsupervised learning, we develop biologically more plausible variants of the restricted Boltzmann machine (RBM), and benchmark their performance on MNIST. We show that RBMs with asymmetric connectivity can still be successfully trained with contrastive divergence, even if no two units are reciprocally connected. Furthermore, RBMs are able to learn if the forward, visible-to-hidden layer weights are kept constant and only the backward, hidden-to-visible layer weights are updated. These findings indicate that neural networks with biologically plausible connectivity support contrastive learning.
Open AccessBiologically plausible unsupervised learning in neural networks with sparse and asymmetric connectivity
In the search for biologically plausible but mathematically precise theories of learning in the brain, recent studies have begun to investigate how key assumptions underlying algorithms for supervised learning in artificial neural networks can be relaxed in biologically plausible ways. Turning to unsupervised learning, we develop biologically more plausible variants of the restricted Boltzmann machine (RBM), and benchmark their performance on MNIST. We show that RBMs with asymmetric connectivity can still be successfully trained with contrastive divergence, even if no two units are reciprocally connected. Furthermore, RBMs are able to learn if the forward, visible-to-hidden layer weights are kept constant and only the backward, hidden-to-visible layer weights are updated. These findings indicate that neural networks with biologically plausible connectivity support contrastive learning.
- Open Access
Somnotate: A probabilistic sleep stage classifier for studying vigilance state transitions
Electrophysiological recordings from freely behaving animals are a widespread and powerful mode of investigation in sleep research. These recordings generate large amounts of data that require sleep stage annotation (polysomnography), in which the data is parcellated according to three vigilance states: awake, rapid eye movement (REM) sleep, and non-REM (NREM) sleep. Manual and computational annotation methods currently ignore intermediate states because the classification features become ambiguous. However, these intermediate states contain important information regarding vigilance state dynamics. Here, we present a new classifier, “Somnotate”, which produces automated annotation accuracies that exceed human expert performance on mouse electrophysiological data, is robust to errors in the training data, compatible with different recording configurations, and maintains high performance during experimental interventions. Somnotate is a probabilistic classifier based on a combination of linear discriminant analysis (LDA) with a hidden Markov model (HMM). A unique feature of Somnotate is that it quantifies and reports the certainty of its annotations, enabling the experimenter to identify ambiguous recording periods in a principled manner. We leverage this feature to identify epochs that exhibit intermediate vigilance states, revealing that many of these cluster around state transitions, whereas others correspond to failed attempts to transition. We show that the success rates of different transitions can be experimentally manipulated and explain previously observed sleep patterns. Somnotate can thus facilitate the study of sleep stage transitions and offers new insight into the mechanisms underlying sleep-wake dynamics. Author summary: Typically, the three different vigilance states – awake, REM sleep, and non-REM sleep – exhibit distinct features that are readily recognised in electrophysiological recordings. However, particularly around vigilance state transitions, epochs often exhibit features from more than one state. These intermediate vigilance states pose challenges for existing manual and automated classification methods, and are hence often ignored. Here, we present ‘Somnotate’ - an open-source, highly accurate and robust sleep stage classifier, which supports research into intermediate states and sleep stage dynamics. Somnotate quantifies and reports the certainty of its annotations, enabling the experimenter to identify abnormal epochs in a principled manner. We use this feature to identify intermediate states and to detect unsuccessful attempts to switch between vigilance states. This provides new insights into the mechanisms of vigilance state transitions in mice, and creates new opportunities for future experiments.
Open AccessSomnotate: A probabilistic sleep stage classifier for studying vigilance state transitions
Electrophysiological recordings from freely behaving animals are a widespread and powerful mode of investigation in sleep research. These recordings generate large amounts of data that require sleep stage annotation (polysomnography), in which the data is parcellated according to three vigilance states: awake, rapid eye movement (REM) sleep, and non-REM (NREM) sleep. Manual and computational annotation methods currently ignore intermediate states because the classification features become ambiguous. However, these intermediate states contain important information regarding vigilance state dynamics. Here, we present a new classifier, “Somnotate”, which produces automated annotation accuracies that exceed human expert performance on mouse electrophysiological data, is robust to errors in the training data, compatible with different recording configurations, and maintains high performance during experimental interventions. Somnotate is a probabilistic classifier based on a combination of linear discriminant analysis (LDA) with a hidden Markov model (HMM). A unique feature of Somnotate is that it quantifies and reports the certainty of its annotations, enabling the experimenter to identify ambiguous recording periods in a principled manner. We leverage this feature to identify epochs that exhibit intermediate vigilance states, revealing that many of these cluster around state transitions, whereas others correspond to failed attempts to transition. We show that the success rates of different transitions can be experimentally manipulated and explain previously observed sleep patterns. Somnotate can thus facilitate the study of sleep stage transitions and offers new insight into the mechanisms underlying sleep-wake dynamics. Author summary: Typically, the three different vigilance states – awake, REM sleep, and non-REM sleep – exhibit distinct features that are readily recognised in electrophysiological recordings. However, particularly around vigilance state transitions, epochs often exhibit features from more than one state. These intermediate vigilance states pose challenges for existing manual and automated classification methods, and are hence often ignored. Here, we present ‘Somnotate’ - an open-source, highly accurate and robust sleep stage classifier, which supports research into intermediate states and sleep stage dynamics. Somnotate quantifies and reports the certainty of its annotations, enabling the experimenter to identify abnormal epochs in a principled manner. We use this feature to identify intermediate states and to detect unsuccessful attempts to switch between vigilance states. This provides new insights into the mechanisms of vigilance state transitions in mice, and creates new opportunities for future experiments.