James P. Roach

Memory recall and spike-frequency adaptation

The brain can reproduce memories from partial data; this ability is critical for memory recall. The process of memory recall has been studied using autoassociative networks such as the Hopfield model. This kind of model reliably converges to stored patterns that contain the memory. However, it is unclear how the behavior is controlled by the brain so that after convergence to one configuration, it can proceed with recognition of another one. In the Hopfield model, this happens only through unrealistic changes of an effective global temperature that destabilizes all stored configurations. Here we show that spike-frequency adaptation (SFA), a common mechanism affecting neuron activation in the brain, can provide state-dependent control of pattern retrieval. We demonstrate this in a Hopfield network modified to include SFA, and also in a model network of biophysical neurons. In both cases, SFA allows for selective stabilization of attractors with different basins of attraction, and also for temporal dynamics of attractor switching that is not possible in standard autoassociative schemes. The dynamics of our models give a plausible account of different sorts of memory retrieval.

Connexin 43 gap junctions contribute to brain endothelial barrier hyperpermeability in familial cerebral cavernous malformations type III by modulating tight junction structure

Familial cerebral cavernous malformations type III (fCCM3) is a disease of the cerebrovascular system caused by loss‐of‐function mutations in ccm3 that result in dilated capillary beds that are susceptible to hemorrhage. Before hemorrhage, fCCM3 lesions are characterized by a hyperpermeable blood‐brain barrier (BBB), the key pathologic feature of fCCM3. We demonstrate that connexin 43 (Cx43), a gap junction (GJ) protein that is incorporated into the BBB junction complex, is up‐regulated in lesions of a murine model of f CCM3. Small interfering RNA‐mediated ccm3 knockdown (CCM3KD) in brain endothelial cells in vitro increased Cx43 protein expression, GJ plaque size, GJ intracellular communication (GJIC), and barrier permeability. CCM3KD hyperpermeability was rescued by GAP27, a peptide gap junction and hemichannel inhibitor of Cx43 GJIC. Tight junction (TJ) protein, zonula occludens 1 (ZO‐1), accumulated at Cx43 GJs in CCM3KD cells and displayed fragmented staining at TJs. The GAP27‐mediated inhibition of Cx43 GJs in CCM3KD cells restored ZO‐1 to TJ structures and reduced plaque accumulation at Cx43 GJs. The TJ protein, Claudin‐5, was also fragmented at TJs in CCM3KD cells, and GAP27 treatment lengthened TJ‐associated fragments and increased Claudin 5‐Claudin 5 transinteraction. Overall, we demonstrate that Cx43 GJs are aberrantly increased in fCCM3 and regulate barrier permeability by a TJ‐dependent mechanism.—Johnson, A. M., Roach, J. P., Hu, A., Stamatovic, S. M., Zochowski, M. R., Keep, R. F., Andjelkovic, A. V. Connexin 43 gap junctions contribute to brain endothelial barrier hyperpermeability in familial cerebral cavernous malformations type III by modulating tight junction structure. FASEB J. 32, 2615–2629 (2018). www.fasebj.org

Formation and Dynamics of Waves in a Cortical Model of Cholinergic Modulation.

Acetylcholine (ACh) is a regulator of neural excitability and one of the neurochemical substrates of sleep. Amongst the cellular effects induced by cholinergic modulation are a reduction in spike-frequency adaptation (SFA) and a shift in the phase response curve (PRC). We demonstrate in a biophysical model how changes in neural excitability and network structure interact to create three distinct functional regimes: localized asynchronous, traveling asynchronous, and traveling synchronous. Our results qualitatively match those observed experimentally. Cortical activity during slow wave sleep (SWS) differs from that during REM sleep or waking states. During SWS there are traveling patterns of activity in the cortex; in other states stationary patterns occur. Our model is a network composed of Hodgkin-Huxley type neurons with a M-current regulated by ACh. Regulation of ACh level can account for dynamical changes between functional regimes. Reduction of the magnitude of this current recreates the reduction in SFA the shift from a type 2 to a type 1 PRC observed in the presence of ACh. When SFA is minimal (in waking or REM sleep state, high ACh) patterns of activity are localized and easily pinned by network inhomogeneities. When SFA is present (decreasing ACh), traveling waves of activity naturally arise. A further decrease in ACh leads to a high degree of synchrony within traveling waves. We also show that the level of ACh determines how sensitive network activity is to synaptic heterogeneity. These regimes may have a profound functional significance as stationary patterns may play a role in the proper encoding of external input as memory and traveling waves could lead to synaptic regularization, giving unique insights into the role and significance of ACh in determining patterns of cortical activity and functional differences arising from the patterns.

Resonance with subthreshold oscillatory drive organizes activity and optimizes learning in neural networks

Significance Networks of neurons need to reliably encode and replay patterns and sequences of activity. In the brain, sequences of spatially coding neurons are replayed in both the forward and reverse direction in time with respect to their order in recent experience. As of yet there is no network-level or biophysical mechanism known that can produce both modes of replay within the same network. Here we propose that resonance, a property of neurons, paired with subthreshold oscillations in neural input facilitate network-level learning of fixed and sequential activity patterns and lead to both forward and reverse replay. Network oscillations across and within brain areas are critical for learning and performance of memory tasks. While a large amount of work has focused on the generation of neural oscillations, their effect on neuronal populations’ spiking activity and information encoding is less known. Here, we use computational modeling to demonstrate that a shift in resonance responses can interact with oscillating input to ensure that networks of neurons properly encode new information represented in external inputs to the weights of recurrent synaptic connections. Using a neuronal network model, we find that due to an input current-dependent shift in their resonance response, individual neurons in a network will arrange their phases of firing to represent varying strengths of their respective inputs. As networks encode information, neurons fire more synchronously, and this effect limits the extent to which further “learning” (in the form of changes in synaptic strength) can occur. We also demonstrate that sequential patterns of neuronal firing can be accurately stored in the network; these sequences are later reproduced without external input (in the context of subthreshold oscillations) in both the forward and reverse directions (as has been observed following learning in vivo). To test whether a similar mechanism could act in vivo, we show that periodic stimulation of hippocampal neurons coordinates network activity and functional connectivity in a frequency-dependent manner. We conclude that resonance with subthreshold oscillations provides a plausible network-level mechanism to accurately encode and retrieve information without overstrengthening connections between neurons.

Sub-threshold resonance organizes activity and optimizes learning in neural networks.

Network oscillations across and within brain areas are critical for learning and performance in memory tasks. While a large amount of work has focused on the generation of neural oscillations, their effects on neuronal populations’ spiking activity and information encoding is less known. Here, we use computational modeling and in vivo recording to demonstrate that a shift in sub-threshold resonance can interact with oscillating input to ensure that networks of neurons properly encode new information represented in external inputs to the weights of recurrent synaptic connections. Using a neuronal network model, we find that due to an input-current dependent shift in their resonance response, individual neurons in a network will arrange their phases of firing to represent varying strengths of their respective inputs. As networks encode information, neurons fire more synchronously, and this effect limits the extent to which further “learning” (in the form of changes in synaptic strength) can occur. We also demonstrate that sequential patterns of neuronal firing can be accurately stored in the network; these sequences are later reproduced without external input (in the context of sub-threshold oscillations) in both the forward and reverse directions (as has been observed following learning in vivo). To test whether a similar mechanism could act in vivo, we show that periodic stimulation of hippocampal neurons coordinates network activity and functional connectivity in a frequency-dependent manner. We conclude that sub-threshold resonance provides a plausible network-level mechanism to accurately encode and retrieve information without over-strengthening connections between neurons.

Modifications in Network Structure and Excitability May Drive Differential Activity-Dependent Integration of Granule Cells into Dentate Gyrus Circuits During Normal and Pathological Adult Neurogenesis

The granule cells born during adult neurogenesis in the dentate gyrus (DG) are thought to be involved in formation of new memory representations. At the same time, they are implicated in exacerbation of the pathology during epilepsy. Moreover, it has been found that patterns of their integration into DG circuits are significantly different in health and pathology. The aim of this contribution is to identify network-wide structural and dynamical mechanisms underlying this differential incorporation of the newly born cells, as well as resulting changes in activity patterns in the network. We show that, on the one hand, decreased network-wide inhibition and long-range excitatory connectivity alone can result in significant changes in augmentation patterns of new cells, such as increased survival rate of new cells, emergence of globally synchronized activity patterns, and decreased correlation between network drive and location of the surviving cells. On the other hand, we show that changes in excitability, namely phase response curves, of newly born cells can also lead to emergence of globally coherent activity patterns that are not responsive to local input properties. These results indicate that both of these mechanisms can be responsible for reorganization of neuronal pathological integration during adult neurogenesis.

Modeling the formation and dynamics of cortical waves induced by cholinergic modulation

States of arousal, or consciousness with the brain are regulated largely by the neurotransmitter acetylcholine (ACh). Specifically, ACh is likely responsible for the transition between slow wave sleep (SWS; where ACh is absent) and rapid eye movement sleep or waking states (where ACh is high). Patterns of neural activity within the cerebral cortex corresponding to these states are markedly different. During SWS there are traveling waves of intense activity in the cortex while in other states locally organized stationary patterns occur [1]. From a functional perspective, stationary patterns are likely to be important for working memory and attention dynamics while traveling waves could lead to synaptic renormalization [2]. The mechanism for how changes on the cellular level are translated to patterns on the network level is not understood. In this work we give a model for the action of ACh on a network of neurons of the Hodgkin-Huxley type with a current that is regulated by ACh that induces spike-frequency adaptation (SFA) [3]. The cells are coupled in a center-surround scheme. When SFA is minimal (such as in waking or REM sleep state, high ACh) patterns of activity are localized and easily pinned to regions defined by enhanced recurrent excitation. Increasing the level SFA is present (by increasing ACh), traveling waves of activity naturally arise. Depending on the strength of inhibitory coupling within the network, SFA is able to induce a wide variety of dynamical regimes (Figure ​(Figure1).1). We present a detailed mechanism that shows that the level of inhibition sets the spatial extent of network activity and that SFA defines the temporal scope, which is directly modulated by ACh in the model. These model calculations give unique insights into the role and significance of ACh in determining patterns of cortical activity and functional differences arising from these patterns. Figure 1 An illustration of the dynamics sampled by scannig inhibitory strength,(wi e), and gKs . In this model gKs is increased to simulate decreasing ACh levels. In a general sense, the spatial scope of activity is determined by the excitatory/ inhibitory balance, ...

The interplay of intrinsic excitability and network topology in spatiotemporal pattern generation in neural networks

It is clear that spatiotemporal patterning in brain networks is a complex outcome of network physical connectivity and dynamical properties of interacting neurons, however characterization of this interaction remains elusive. These dynamical properties of the cells are affected/controlled by various neuromodulators secreted by the brain at various cognitive cycles or as a part of the response to the incoming stimuli. During sleep the brain cycles though distinct spatiotemporal patterns of neural activity. Acetylcholine (ACh) is a major regulatory factor of sleep states and plays an important role in the transition from slow wave sleep to waking or rapid eye movement sleep. Slow wave sleep is a slow oscillation in firing rate that travels through the cortical network and occurs when ACh levels are low [1]. At the cellular level, ACh causes changes in neural excitability by shifting the neural phase response curve (PRC)

Network topology and intrinsic excitability of the existing network drive integration patterns in a model of adult neurogenesis

Neurogenesis in the adult hippocampus is critical process in learning and memory where immature granule cells project dendrites to the molecular layer and axons to CA3 pyramidal cells[1]. While the developmental processes that govern the maturation of new granule cells have been well characterized, the manner in which the structure of the established network and the intrinsic excitability of the neurons within that network govern the incorporation of new neurons into the network is still an unanswered question. The formation of inward and outward connections is the most obvious characteristic of integration of maturing neurons. Thus expanding the understanding of how shifts the timing of a postsynaptic spike after presynaptic spikes, which is measured by the phase response curve (PRC), changes the wiring of new cells into the network is an important line of research. Our group has previously used a leaky integrate and fire (LIF) computational model to investigate how new neurons integrate into networks of various topologies and activity levels. A limitation of this study is that the PRC of LIF model neurons is type 1, or purely excitatory, while the PRC hippocampal CA3 cells have been shown to be type 2, or both excitatory and inhibitory[2]. To address this, we will perform simulations of networks using Hodgkin-Huxley model neurons, which have a PRC that can be switched from type 2 to type 1 by decreasing the conductance of a M-type slow voltage dependant potassium current[3]. The networks are laid upon a 2D lattice with external stimulation at the centermost cells. Networks composed of type 2 neurons display bursting dynamics with broad activity, while those composed of type 1 neurons have activity clustered around the stimulus and low levels of synchrony. These differences in network dynamics will change how newborn neurons form connections with the existing network. Using a factorial experimental design the incorporation newborn cells will be investigated in four ways; type 1 cells added to a type 1 network, type 1 cells added to a type 2 network, type 2 cells added to a type 1 network, and type 2 cells added to a type 2 network. When both new and mature cells are type 1, 24% of new cells reached an activity level high enough for survival. In the case of type 2 cells, all introduced survived. The results of these simulations for both PRC types will allow for comparison with our previous work and show how changes in neuronal excitability.