MRC Brain Network Dynamics Unit

Adaptive deep brain stimulation uses feedback about the state of neural circuits to control stimulation rather than delivering fixed stimulation all the time, as currently performed. In patients with Parkinson's disease, elevations in beta activity (13-35 Hz) in the subthalamic nucleus have been demonstrated to correlate with clinical impairment and have provided the basis for feedback control in trials of adaptive deep brain stimulation. These pilot studies have suggested that adaptive deep brain stimulation may potentially be more effective, efficient and selective than conventional deep brain stimulation, implying mechanistic differences between the two approaches. Here we test the hypothesis that such differences arise through differential effects on the temporal dynamics of beta activity. The latter is not constantly increased in Parkinson's disease, but comes in bursts of different durations and amplitudes. We demonstrate that the amplitude of beta activity in the subthalamic nucleus increases in proportion to burst duration, consistent with progressively increasing synchronization. Effective adaptive deep brain stimulation truncated long beta bursts shifting the distribution of burst duration away from long duration with large amplitude towards short duration, lower amplitude bursts. Critically, bursts with shorter duration are negatively and bursts with longer duration positively correlated with the motor impairment off stimulation. Conventional deep brain stimulation did not change the distribution of burst durations. Although both adaptive and conventional deep brain stimulation suppressed mean beta activity amplitude compared to the unstimulated state, this was achieved by a selective effect on burst duration during adaptive deep brain stimulation, whereas conventional deep brain stimulation globally suppressed beta activity. We posit that the relatively selective effect of adaptive deep brain stimulation provides a rationale for why this approach could be more efficacious than conventional continuous deep brain stimulation in the treatment of Parkinson's disease, and helps inform how adaptive deep brain stimulation might best be delivered.

Exaggerated basal ganglia beta activity (13-35 Hz) is commonly found in patients with Parkinson's disease and can be suppressed by dopaminergic medication, with the degree of suppression being correlated with the improvement in motor symptoms. Importantly, beta activity is not continuously elevated, but fluctuates to give beta bursts. The percentage number of longer beta bursts in a given interval is positively correlated with clinical impairment in Parkinson's disease patients. Here we determine whether the characteristics of beta bursts are dependent on dopaminergic state. Local field potentials were recorded from the subthalamic nucleus of eight Parkinson's disease patients during temporary lead externalization during surgery for deep brain stimulation. The recordings took place with the patient quietly seated following overnight withdrawal of levodopa and after administration of levodopa. Beta bursts were defined by applying a common amplitude threshold and burst characteristics were compared between the two drug conditions. The amplitude of beta bursts, indicative of the degree of local neural synchronization, progressively increased with burst duration. Treatment with levodopa limited this evolution leading to a relative increase of shorter, lower amplitude bursts. Synchronization, however, was not limited to local neural populations during bursts, but also, when such bursts were cotemporaneous across the hemispheres, was evidenced by bilateral phase synchronization. The probability of beta bursts and the proportion of cotemporaneous bursts were reduced by levodopa. The percentage number of longer beta bursts in a given interval was positively related to motor impairment, while the opposite was true for the percentage number of short duration beta bursts. Importantly, the decrease in burst duration was also correlated with the motor improvement. In conclusion, we demonstrate that long duration beta bursts are associated with an increase in local and interhemispheric synchronization. This may compromise information coding capacity and thereby motor processing. Dopaminergic activity limits this uncontrolled beta synchronization by terminating long duration beta bursts, with positive consequences on network state and motor symptoms.

This review article summarises recently proposed theories on how neural circuits in the brain could approximate the error back-propagation algorithm used by artificial neural networks. Computational models implementing these theories achieve learning as efficient as artificial neural networks, but they use simple synaptic plasticity rules based on activity of presynaptic and postsynaptic neurons. The models have similarities, such as including both feedforward and feedback connections, allowing information about error to propagate throughout the network. Furthermore, they incorporate experimental evidence on neural connectivity, responses, and plasticity. These models provide insights on how brain networks might be organised such that modification of synaptic weights on multiple levels of cortical hierarchy leads to improved performance on tasks.

The brain is capable of producing coordinated fast changing neural dynamics across multiple brain regions in order to adapt to rapidly changing environments. However, it is non-trivial to identify multiregion dynamics at fast sub-second time-scales in electrophysiological data. We propose a method that, with no knowledge of any task timings, can simultaneously identify and describe fast transient multiregion dynamics in terms of their temporal, spectral and spatial properties. The approach models brain activity using a discrete set of sequential states, with each state distinguished by its own multiregion spectral properties. This can identify potentially very short-lived visits to a brain state, at the same time as inferring the state's properties, by pooling over many repeated visits to that state. We show how this can be used to compute state-specific measures such as power spectra and coherence. We demonstrate that this can be used to identify short-lived transient brain states with distinct power and functional connectivity (e.g., coherence) properties in an MEG data set collected during a volitional motor task.

INTRODUCTION & OBJECTIVES: Adaptive deep brain stimulation (aDBS) uses feedback from brain signals to guide stimulation. A recent acute trial of unilateral aDBS showed that aDBS can lead to substantial improvements in contralateral hemibody Unified Parkinson's Disease Rating Scale (UPDRS) motor scores and may be superior to conventional continuous DBS in Parkinson's disease (PD). We test whether potential benefits are retained with bilateral aDBS and in the face of concurrent medication. METHODS: We applied bilateral aDBS in 4 patients with PD undergoing DBS of the subthalamic nucleus. aDBS was delivered bilaterally with independent triggering of stimulation according to the amplitude of β activity at the corresponding electrode. Mean stimulation voltage was 3.0±0.1 volts. Motor assessments consisted of double-blinded video-taped motor UPDRS scores that included both limb and axial features. RESULTS: UPDRS scores were 43% (p=0.04; Cohen's d=1.62) better with aDBS than without stimulation. Motor improvement with aDBS occurred despite an average time on stimulation (ToS) of only 45%. Levodopa was well tolerated during aDBS and led to further reductions in ToS. CONCLUSION: Bilateral aDBS can improve both axial and limb symptoms and can track the need for stimulation across drug states.

The ability to reinstate neuronal assemblies representing mnemonic information is thought to require their consolidation through offline reactivation during sleep/rest. To test this, we detected cell assembly patterns formed by repeated neuronal co-activations in the mouse hippocampus during exploration of spatial environments. We found that the reinstatement of assembly patterns representing a novel, but not a familiar, environment correlated with their offline reactivation and was impaired by closed-loop optogenetic disruption of sharp wave-ripple oscillations. Moreover, we discovered that reactivation was only required for the reinstatement of assembly patterns whose expression was gradually strengthened during encoding of a novel place. The context-dependent reinstatement of assembly patterns whose expression did not gain in strength beyond the first few minutes of spatial encoding was not dependent on reactivation. This demonstrates that the hippocampus can hold concurrent representations of space that markedly differ in their encoding dynamics and their dependence on offline reactivation for consolidation.Video AbstracteyJraWQiOiI4ZjUxYWNhY2IzYjhiNjNlNzFlYmIzYWFmYTU5NmZmYyIsImFsZyI6IlJTMjU2In0.eyJzdWIiOiI2NGM2MzA3ZWQ0NjY1ODI1YTEzZWMxNjZmOTM0YTcwNiIsImtpZCI6IjhmNTFhY2FjYjNiOGI2M2U3MWViYjNhYWZhNTk2ZmZjIiwiZXhwIjoxNjc5NTI1NjQ3fQ.qMBzoMWhkGM9Px-KMy6xUpszcRfk4Vk9ovilJTZh22a8igEB3m5SfM2HUquK7wFP1xZC8dtiID8rCIFXLKPQiiGgi7JU3epoW2lPOw4eGoCieobYEkVblYX-D4fTo71FMEAd0KC-4MRUR-dgCG-xOLTA3v0Budd4Z_NqklcH1LBmZcNjfGKmmb26Y-EzwvTud35QdILLeQU2QrL-Ls2qtWXY0UZM3ufKv_n-F9SiCbh_8c37l_lrAynXR18p898UKBb37Z-1LPldHLRMjczLZUAsmE9eWOt-heSTWCNRiB-H8Q9CSd-A4lGAtrT9hNCf4Z6BowWZAoQjHL09gAAbfg(mp4, (30.51 MB) Download video

OBJECTIVE: Beta band oscillations in the subthalamic nucleus (STN) have been proposed as a pathophysiological signature in patients with Parkinson's disease (PD). The aim of this study was to investigate the potential association between oscillatory activity in the STN and symptom severity in PD. METHODS: Subthalamic local field potentials were recorded from 63 PD patients in a dopaminergic OFF state. Power-spectra were analyzed for the frequency range from 5 to 95 Hz and correlated with individual UPDRS-III motor scores in the OFF state. RESULTS: A correlation between total UPDRS-III scores and 8 to 35 Hz activity was revealed across all patients (ρ = 0.44, P < .0001). When correlating each frequency bin, a narrow range from 10 to 15 Hz remained significant for the correlation (false discovery rate corrected P < .05). CONCLUSION: Our results show a correlation between local STN 8 to 35 Hz power and impairment in PD, further supporting the role of subthalamic oscillatory activity as a potential biomarker for PD.

Chronic dopamine depletion in Parkinson's disease leads to progressive motor and cognitive impairment, which is associated with the emergence of characteristic patterns of synchronous oscillatory activity within cortico-basal-ganglia circuits. Deep brain stimulation of the subthalamic nucleus is an effective treatment for Parkinson's disease, but its influence on synchronous activity in cortico-basal-ganglia loops remains to be fully characterized. Here, we demonstrate that deep brain stimulation selectively suppresses certain spatially and spectrally segregated resting state subthalamic nucleus-cortical networks. To this end we used a validated and novel approach for performing simultaneous recordings of the subthalamic nucleus and cortex using magnetoencephalography (during concurrent subthalamic nucleus deep brain stimulation). Our results highlight that clinically effective subthalamic nucleus deep brain stimulation suppresses synchrony locally within the subthalamic nucleus in the low beta oscillatory range and furthermore that the degree of this suppression correlates with clinical motor improvement. Moreover, deep brain stimulation relatively selectively suppressed synchronization of activity between the subthalamic nucleus and mesial premotor regions, including the supplementary motor areas. These mesial premotor regions were predominantly coupled to the subthalamic nucleus in the high beta frequency range, but the degree of deep brain stimulation-associated suppression in their coupling to the subthalamic nucleus was not found to correlate with motor improvement. Beta band coupling between the subthalamic nucleus and lateral motor areas was not influenced by deep brain stimulation. Motor cortical coupling with subthalamic nucleus predominantly involved driving of the subthalamic nucleus, with those drives in the higher beta frequency band having much shorter net delays to subthalamic nucleus than those in the lower beta band. These observations raise the possibility that cortical connectivity with the subthalamic nucleus in the high and low beta bands may reflect coupling mediated predominantly by the hyperdirect and indirect pathways to subthalamic nucleus, respectively, and that subthalamic nucleus deep brain stimulation predominantly suppresses the former. Yet only the change in strength of local subthalamic nucleus oscillations correlates with the degree of improvement during deep brain stimulation, compatible with the current view that a strengthened hyperdirect pathway is a prerequisite for locally generated beta activity but that it is the severity of the latter that may determine or index motor impairment.

This paper provides an easy to follow tutorial on the free-energy framework for modelling perception developed by Friston, which extends the predictive coding model of Rao and Ballard. These models assume that the sensory cortex infers the most likely values of attributes or features of sensory stimuli from the noisy inputs encoding the stimuli. Remarkably, these models describe how this inference could be implemented in a network of very simple computational elements, suggesting that this inference could be performed by biological networks of neurons. Furthermore, learning about the parameters describing the features and their uncertainty is implemented in these models by simple rules of synaptic plasticity based on Hebbian learning. This tutorial introduces the free-energy framework using very simple examples, and provides step-by-step derivations of the model. It also discusses in more detail how the model could be implemented in biological neural circuits. In particular, it presents an extended version of the model in which the neurons only sum their inputs, and synaptic plasticity only depends on activity of pre-synaptic and post-synaptic neurons.

Adaptation to the ever-changing world is critical for survival, and our brains are particularly tuned to remember events that differ from previous experiences. Novel experiences induce dopamine release in the hippocampus, a process which promotes memory persistence. While axons from the ventral tegmental area (VTA) were generally thought to be the exclusive source of hippocampal dopamine, recent studies have demonstrated that noradrenergic neurons in the locus coeruleus (LC) corelease noradrenaline and dopamine in the hippocampus and that their dopamine release boosts memory retention as well. In this opinion article, we propose that the projections originating from the VTA and the LC belong to two distinct systems that enhance memory of novel events. Novel experiences that share some commonality with past ones ('common novelty') activate the VTA and promote semantic memory formation via systems memory consolidation. By contrast, experiences that bear only a minimal relationship to past experiences ('distinct novelty') activate the LC to trigger strong initial memory consolidation in the hippocampus, resulting in vivid and long-lasting episodic memories.

Sleep is crucial for healthy cognition, including memory. The two main phases of sleep, REM (rapid eye movement) and non-REM sleep, are associated with characteristic electrophysiological patterns that are recorded using surface and intracranial electrodes. These patterns include sharp-wave ripples, cortical slow oscillations, delta waves, and spindles during non-REM sleep and theta oscillations during REM sleep. They reflect the precisely timed activity of underlying neural circuits. Here, we review how these electrical signatures have been guiding our understanding of the circuits and processes sustaining memory consolidation during sleep, focusing on hippocampal theta oscillations and sharp-wave ripples and how they coordinate with cortical patterns. Finally, we highlight how these brain patterns could also sustain sleep-dependent homeostatic processes and evoke several potential future directions for research on the memory function of sleep.

The subthalamic nucleus (STN) of the basal ganglia appears to have a potent role in action and cognition. Anatomical and imaging studies show that different frontal cortical areas directly project to the STN via so-called hyperdirect pathways. This review reports some of the latest findings about such circuits, including simultaneous recordings from cortex and the STN in humans, single-unit recordings in humans, high-resolution fMRI, and neurocomputational modeling. We argue that a major function of the STN is to broadly pause behavior and cognition when stop signals, conflict signals, or surprise signals occur, and that the fronto-STN circuits for doing this, at least for stopping and conflict, are dissociable anatomically and in terms of their spectral reactivity. We also highlight recent evidence for synchronization of oscillations between prefrontal cortex and the STN, which may provide a preferential "window in time" for single neuron communication via long-range connections.

This paper revisits the dynamic causal modelling of fMRI timeseries by replacing the usual (Taylor) approximation to neuronal dynamics with a neural mass model of the canonical microcircuit. This provides a generative or dynamic causal model of laminar specific responses that can generate haemodynamic and electrophysiological measurements. In principle, this allows the fusion of haemodynamic and (event related or induced) electrophysiological responses. Furthermore, it enables Bayesian model comparison of competing hypotheses about physiologically plausible synaptic effects; for example, does attentional modulation act on superficial or deep pyramidal cells - or both? In this technical note, we describe the resulting dynamic causal model and provide an illustrative application to the attention to visual motion dataset used in previous papers. Our focus here is on how to answer long-standing questions in fMRI; for example, do haemodynamic responses reflect extrinsic (afferent) input from distant cortical regions, or do they reflect intrinsic (recurrent) neuronal activity? To what extent do inhibitory interneurons contribute to neurovascular coupling? What is the relationship between haemodynamic responses and the frequency of induced neuronal activity? This paper does not pretend to answer these questions; rather it shows how they can be addressed using neural mass models of fMRI timeseries.

Beta oscillations are a dominant feature of the sensorimotor system. A transient and prominent increase in beta oscillations is consistently observed across the sensorimotor cortical-basal ganglia network after cessation of voluntary movement: the post-movement beta synchronization (PMBS). Current theories about the function of the PMBS have been focused on either the closure of motor response or the processing of sensory afferance. Computational models of sensorimotor control have emphasized the importance of the integration between feedforward estimation and sensory feedback, and therefore the putative motor and sensory functions of beta oscillations may reciprocally interact with each other and in fact be indissociable. Here we show that the amplitude of sensorimotor PMBS is modulated by the history of visual feedback of task-relevant errors, and negatively correlated with the trial-to-trial exploratory adjustment in a sensorimotor adaptation task in young healthy human subjects. The PMBS also negatively correlated with the uncertainty associated with the feedforward estimation, which was recursively updated in light of new sensory feedback, as identified by a Bayesian learning model. These results reconcile the two opposing motor and sensory views of the function of PMBS, and suggest a unifying theory in which PMBS indexes the confidence in internal feedforward estimation in Bayesian sensorimotor integration. Its amplitude simultaneously reflects cortical sensory processing and signals the need for maintenance or adaptation of the motor output, and if necessary, exploration to identify an altered sensorimotor transformation. SIGNIFICANCE STATEMENT: For optimal sensorimotor control, sensory feedback and feedforward estimation of a movement's sensory consequences should be weighted by the inverse of their corresponding uncertainties, which require recursive updating in a dynamic environment. We show that post-movement beta activity (13-30 Hz) over sensorimotor cortex in young healthy subjects indexes the evaluation of uncertainty in feedforward estimation. Our work contributes to the understanding of the function of beta oscillations in sensorimotor control, and provides further insight into how aberrant beta activity can contribute to the pathophysiology of movement disorders.

To efficiently learn from feedback, cortical networks need to update synaptic weights on multiple levels of cortical hierarchy. An effective and well-known algorithm for computing such changes in synaptic weights is the error backpropagation algorithm. However, in this algorithm, the change in synaptic weights is a complex function of weights and activities of neurons not directly connected with the synapse being modified, whereas the changes in biological synapses are determined only by the activity of presynaptic and postsynaptic neurons. Several models have been proposed that approximate the backpropagation algorithm with local synaptic plasticity, but these models require complex external control over the network or relatively complex plasticity rules. Here we show that a network developed in the predictive coding framework can efficiently perform supervised learning fully autonomously, employing only simple local Hebbian plasticity. Furthermore, for certain parameters, the weight change in the predictive coding model converges to that of the backpropagation algorithm. This suggests that it is possible for cortical networks with simple Hebbian synaptic plasticity to implement efficient learning algorithms in which synapses in areas on multiple levels of hierarchy are modified to minimize the error on the output.

SEE MOLL AND ENGEL DOI101093/AWW308 FOR A SCIENTIFIC COMMENTARY ON THIS ARTICLE: Brain regions dynamically engage and disengage with one another to execute everyday actions from movement to decision making. Pathologies such as Parkinson's disease and tremor emerge when brain regions controlling movement cannot readily decouple, compromising motor function. Here, we propose a novel stimulation strategy that selectively regulates neural synchrony through phase-specific stimulation. We demonstrate for the first time the therapeutic potential of such a stimulation strategy for the treatment of patients with pathological tremor. Symptom suppression is achieved by delivering stimulation to the ventrolateral thalamus, timed according to the patient's tremor rhythm. Sustained locking of deep brain stimulation to a particular phase of tremor afforded clinically significant tremor relief (up to 87% tremor suppression) in selected patients with essential tremor despite delivering less than half the energy of conventional high frequency stimulation. Phase-specific stimulation efficacy depended on the resonant characteristics of the underlying tremor network. Selective regulation of neural synchrony through phase-locked stimulation has the potential to both increase the efficiency of therapy and to minimize stimulation-induced side effects.

Studies in dopamine-depleted rats indicate that the external globus pallidus (GPe) contains two main types of GABAergic projection cell; so-called "prototypic" and "arkypallidal" neurons. Here, we used correlative anatomical and electrophysiological approaches in rats to determine whether and how this dichotomous organization applies to the dopamine-intact GPe. Prototypic neurons coexpressed the transcription factors Nkx2-1 and Lhx6, comprised approximately two-thirds of all GPe neurons, and were the major GPe cell type innervating the subthalamic nucleus (STN). In contrast, arkypallidal neurons expressed the transcription factor FoxP2, constituted just over one-fourth of GPe neurons, and innervated the striatum but not STN. In anesthetized dopamine-intact rats, molecularly identified prototypic neurons fired at relatively high rates and with high regularity, regardless of brain state (slow-wave activity or spontaneous activation). On average, arkypallidal neurons fired at lower rates and regularities than prototypic neurons, and the two cell types could be further distinguished by the temporal coupling of their firing to ongoing cortical oscillations. Complementing the activity differences observed in vivo, the autonomous firing of identified arkypallidal neurons in vitro was slower and more variable than that of prototypic neurons, which tallied with arkypallidal neurons displaying lower amplitudes of a "persistent" sodium current important for such pacemaking. Arkypallidal neurons also exhibited weaker driven and rebound firing compared with prototypic neurons. In conclusion, our data support the concept that a dichotomous functional organization, as actioned by arkypallidal and prototypic neurons with specialized molecular, structural, and physiological properties, is fundamental to the operations of the dopamine-intact GPe.

Deep brain stimulation (DBS) for Parkinson's disease (PD) is currently limited by costs, partial efficacy and surgical and stimulation-related side effects. This has motivated the development of adaptive DBS (aDBS) whereby stimulation is automatically adjusted according to a neurophysiological biomarker of clinical state, such as β oscillatory activity (12–30 Hz). aDBS has been studied in parkinsonian primates and patients and has been reported to be more energy efficient and effective in alleviating motor symptoms than conventional DBS (cDBS) at matched amplitudes.

Understanding how the human brain gives rise to complex cognitive processes remains one of the biggest challenges of contemporary neuroscience. While invasive recording in animal models can provide insight into neural processes that are conserved across species, our understanding of cognition more broadly relies upon investigation of the human brain itself. There is therefore an imperative to establish non-invasive tools that allow human brain activity to be measured at high spatial and temporal resolution. In recent years, various attempts have been made to refine the coarse signal available in functional magnetic resonance imaging (fMRI), providing a means to investigate neural activity at the meso-scale, i.e. at the level of neural populations. The most widely used techniques include repetition suppression and multivariate pattern analysis. Human neuroscience can now use these techniques to investigate how representations are encoded across neural populations and transformed by relevant computations. Here, we review the physiological basis, applications and limitations of fMRI repetition suppression with a brief comparison to multivariate techniques. By doing so, we show how fMRI repetition suppression holds promise as a tool to reveal complex neural mechanisms that underlie human cognitive function.This article is part of the themed issue 'Interpreting BOLD: a dialogue between cognitive and cellular neuroscience'.

The hippocampus is crucial for episodic memory, but it is also involved in online prediction. Evidence suggests that a unitary hippocampal code underlies both episodic memory and predictive processing, yet within a predictive coding framework the hippocampal-neocortical interactions that accompany these two phenomena are distinct and opposing. Namely, during episodic recall, the hippocampus is thought to exert an excitatory influence on the neocortex, to reinstate activity patterns across cortical circuits. This contrasts with empirical and theoretical work on predictive processing, where descending predictions suppress prediction errors to 'explain away' ascending inputs via cortical inhibition. In this hypothesis piece, we attempt to dissolve this previously overlooked dialectic. We consider how the hippocampus may facilitate both prediction and memory, respectively, by inhibiting neocortical prediction errors or increasing their gain. We propose that these distinct processing modes depend upon the neuromodulatory gain (or precision) ascribed to prediction error units. Within this framework, memory recall is cast as arising from fictive prediction errors that furnish training signals to optimise generative models of the world, in the absence of sensory data.