Research
Past Accomplishments
Through its monosynaptic connections, gap junctions, and anatomical projections, the hippocampal network generates a range of rhythms that are not completely understood. We believe that the biophysical elements and mechanisms that generate these rhythms are central to the mnemonic function of the hippocampus. In particular, theta oscillations in the hippocampus are believed to be essential for the formation and retrieval of episodic memories [1]. However, it is not clear how information is processed in the hippocampal formation during theta states. To elucidate this, we simultaneously recorded the activity of several (~100) neurons and local field potentials from multiple layers of the hippocampus and entorhinal cortex (EC) in rats performing spatial tasks. Temporal delays between population activities in successive anatomical stages were much longer (~60 ms) than would be expected based on axon conduction velocities and passive synaptic integration of feed-forward excitatory inputs, suggesting that the temporal windows set by the theta cycles allow for local circuit interactions. The discovery of grid cells in the EC led to the hypothesis that the map of the environment, which had long been attributed to the hippocampus, resides in fact in the EC and that the place fields of hippocampal neurons simply reflect the spatial and temporal convergence of entorhinal input. However, the long delays between EC and hippocampal population activities we observed imply that the hippocampus is not a simple integrator for the EC, and that this hypothesis is somewhat incomplete. Instead, local circuits preserve a considerable degree of computational independence in subdivisions of the entorhinal–hippocampal loop [2,3].
To uncover the mechanisms of local circuit computation, it is essential to understand how individual neurons in a given region participate in information processing [3]. The hippocampal CA1 region is one of the most extensively studied regions of the brain, and most studies tacitly assume that pyramidal neurons in this region represent a homogeneous cell population. However, using high-spatial-resolution silicon probes, we found robust physiological differences along the oriens-radiatum axis in the CA1 pyramidal layer. Compared with their superficial counterparts, deep pyramidal cells fired at higher rates, burst more frequently, and were more strongly subject to modulation by the slow oscillations during sleep. Both deep and superficial pyramidal cells fired preferentially at the trough of theta oscillations during maze exploration, whereas deep pyramidal cells shifted their preferred phase of firing to the peak of theta oscillations during rapid eye movement (REM) sleep. Furthermore, non-shifting cells preferred the trough of gamma waves, although most REM-theta-phase-shifting cells fired at the ascending phase of gamma oscillations during waking. Thus, we showed that CA1 pyramidal cells in adjacent sublayers, which might carry different types of information (e.g., context vs. content, novel vs. familiar, pattern completion vs. separation), could address their targets jointly or differentially depending on brain states and oscillations, thereby forming functionally distinct streams and perhaps preventing interference, especially in memory consolidation processes, during REM sleep [4].
To explain how the brain functions, it is essential to understand the details of its activity, such as firing rate and synapse strength. After a comprehensive analysis of large data sets [5], we found that the firing rates of principal neurons in the entorhinal–hippocampal loop showed a lognormal-like distribution in all brain states. We demonstrated that a highly active minority of neurons, which had high spatial information rates, emitted half of the spikes at all times while a very large number of slowly discharging neurons contributed towards the remaining half. Importantly, firing rates of the same neurons showed high correlations in various brain states and testing conditions as well as in both familiar and novel environments. Skewed firing rates of individual neurons may be a result of the skewed distribution of synaptic weights, which is supported by the observed lognormal distribution of the efficacy of spike transfer from principal neurons to interneurons [6]. Thus, we hypothesized that the persistent skewed distribution of firing rates implies that a preconfigured, highly active minority of neurons dominates information transmission in cortical networks [6,7].
Future Research Goals
We continue to study the mechanisms of memory formation and retrieval in hippocampus–EC circuits. In addition, we are studying how the hippocampus and its cortical and subcortical counterparts, such as the amygdala, basal ganglia, thalamus, and neocortex, cooperate to form and retrieve memory.
Dynamics of functional interactions in the hippocampus–EC system during learning and during sleep
The functional connectivity characterizing a neuronal circuit is dynamic. It has been hypothesized that various network oscillations support transient communication across brain structures. The major excitatory inputs to CA1 arise from CA3 and layer 3 of the EC (EC3). How does a single “receiver” tune to multiple “senders”? Based on computational modeling, it has been hypothesized that theta oscillations separate signals from different input regions by parsing them into different phases of the cycle, thereby preventing interference. Consistent with the prediction from the model, firing of CA3 and EC3 principal neurons occurs at different theta phases during spatial exploration [2]. Furthermore, we found that gamma oscillations (30–140 Hz) in the hippocampus could be divided into three distinct frequency bands (30–50 Hz, 50–90 Hz, and 90–140 Hz) based on their phase–amplitude and phase–phase coupling with theta oscillations [8]. These findings suggest that theta–gamma coupling can support multiple time–scale coordination of neuronal spikes within and across structures [8,9]. However, how the functional interaction between subregions is dynamically controlled in different aspects of learning and behavior is still unknown.
To better appreciate the roles of dynamic interactions during leaning and behavior, we plan to simultaneously record neuronal activity and local field potentials from multiple regions in the hippocampus and EC during hippocampus-dependent spatial memory tasks. We will also compare the dynamic interactions during task performance and sleep. It has been reported that CA1 neuronal activity patterns during waking are spontaneously replayed during subsequent sleep in the absence of external stimuli. This process is hypothesized to be essential for memory consolidation. However, the process through which internally generated replayed information in the hippocampus is transferred to the neocortex, perhaps via the EC, and finally stored as memory is unknown. We will address this question by monitoring neuronal activities from multiple subregions in the hippocampus–EC–neocortex circuitry during task performance and during subsequent sleep. These experiments will offer information that is essential to understand how dynamic coordination of neuronal assemblies across different brain regions enables them to communicate with each other, route behaviorally relevant information, and consolidate memory.
Physiological mechanisms of the subicular complex in information processing
To understand the mechanisms of information processing and the control of information flow in the hippocampus, it is essential to reveal comprehensively the input–output relationship between the hippocampus and related structures. The subiculum, presubiculum, and parasubiculum are intimately connected to the hippocampus and EC. However, to date, little effort has been made to understand the physiological mechanisms by which these structures cooperate with the hippocampus and EC. Anatomical studies suggest the following intriguing but largely unexplored mechanisms:
- (1) The subiculum is a major source of both cortical and subcortical efferent projections from the hippocampal formation, and largely independent populations of intermingled neurons in the subiculum separately project to their target regions. This suggests that the output of the hippocampus is routed to different targets with the assistance of the subiculum [10].
- (2) The presubiculum and parasubiculum receive inputs from the anterior thalamic nuclear complex, where the head direction system resides, suggesting that information convergence of head direction and grid cell systems occurs in these areas.
- (3) The presubiculum projects mainly to EC3 whereas the parasubiculum projects to EC2 and the dentate gyrus, suggesting that functional segregation of the trisynaptic (EC2 → DG → CA3 → CA1) and temporoammonic pathway (EC3 → CA1) may originate from the presubiculum and parasubiculum.
To better understand how the subicular complex cooperates with the hippocampus–EC circuitry in information processing and memory formation, we plan to simultaneously record neuronal activity from the subicular complex and hippocampus–EC circuitry during both hippocampus-dependent spatial memory tasks and sleep. We also plan to examine behavior-dependent information processing and routing in the circuitry. This project will enable a better understanding of the physiological mechanisms underlying neuronal communication in the hippocampus–subicular complex–EC system and will thus help elucidate the functions of the hippocampal formation as a whole in memory and behavior.
How do the hippocampus and its cortical and subcortical counterparts cooperate to form and retrieve memory?
The hippocampus plays a pivotal role in memory formation and retrieval, but little is known of how the hippocampus and its counterparts, such as the amygdala, basal ganglia, thalamus, and neocortex, cooperate in this process. In particular, some reports suggest that newly acquired information is replayed and consolidated during subsequent sleep. It has been hypothesized that interactions between the hippocampus and other structures during sleep are essential for memory formation. However, the mechanism underlying this process in the brain is unclear [11–14], especially in the context of cooperation between brain regions [2,9]. For example, we do not remember all our daily experiences. How is relevant information selected, for example, through interaction between the thalamus and neocortex? Do we erase unnecessary information (e.g., in the neocortex) proactively during sleep to spare the capacity of our memory for future usage? Are good and bad memories processed in distinct manners, perhaps with the aid of the amygdala and basal ganglia? How do the hippocampus and other structures communicate with each other in these processes? Finally, what is the relationship between spontaneous neuronal activity during sleep and neuronal activity during wakefulness in each brain region [6,7,13]?
To address these questions, we will record neuronal activities from the hippocampus, amygdala, basal ganglia, thalamus, and neocortex during various tasks and during sleep before and after the tasks. We will also quantitatively characterize the activities of neurons, local field potentials, interactions between brain regions, and behavioral performance. By doing this, we will gain a better appreciation of how the hippocampus cooperates with other structures to select behaviorally relevant information, consolidate memory, and guide behavior.
Dissection of pathway- and cell-type-specific roles in network dynamics and memory formation
Different types of interneurons and neuromodulatory systems may contribute distinctly to network dynamics and to particular aspects of behavior. Furthermore, principal neurons are not homogeneous even in a given region or layer but show differential molecular expression, form specialized subnetworks, project to different regions, and are functionally heterogeneous [4]. However, the precise roles of distinct neuronal types in network dynamics, local circuit interactions, and information processing in behaving animals are unknown.
To address these questions, we will perform the following experiments using optogenetics in awake and sleeping rodents: (1) using a combination of silicon probe recording and optical stimulation of genetically labeled channelrhodopsin/halorhodopsin-expressing neurons, we will identify the cell types of recorded neurons and monitor their activity during task performance and sleep; (2) we will excite and suppress distinct populations of neurons using various optical stimulation patterns to perturb the network dynamics and temporal evolution of neuronal assemblies in the hippocampus-EC circuitry; and (3) finally, we will investigate the role of neuromodulatory systems in network dynamics and memory. Both waking and REM sleep are associated with theta oscillations in the hippocampus and desynchronized activity in the neocortex, but the firing of serotonergic, catecholaminergic, and histaminergic neurons is largely absent in the latter. By activating these neurons during REM sleep or silencing them during waking, one can understand the contributions of neuromodulatory systems towards distinct network dynamics during waking and REM sleep as well as the discrete roles of these states in memory formation. These experiments will enrich our understanding of pathway- and cell-type-specific roles in the coordination of neuronal activity in local network dynamics and memory formation.
- Mizuseki K, *Buzsaki G (2014)
Theta oscillations decrease spike synchrony in the hippocampus and entorhinal cortex.
Philos. Trans. R. Soc. Lond B Biol. Sci. 369, 20120530. [PubMed] - Mizuseki K, Sirota A, Pastalkova E, *Buzsaki G (2009)
Theta oscillations provide temporal windows for local circuit computation in the entorhinal-hippocampal loop.
Neuron 64, 267-280. [PubMed] - Mizuseki K, Royer S, Diba K, *Buzsaki G (2012)
Activity dynamics and behavioral correlates of CA3 and CA1 hippocampal pyramidal neurons.
Hippocampus 22, 1659-1680. [PubMed] - Mizuseki K, Diba K, Pastalkova E, *Buzsaki G (2011)
Hippocampal CA1 pyramidal cells form functionally distinct sublayers.
Nat. Neurosci. 14, 1174-1181. [PubMed] - *Mizuseki K, Diba K, Pastalkova E, Teeters J, Sirota A, *Buzsáki G (2014)
Neurosharing: Large-scale data sets (spike, LFP) recorded from the hippocampal.htmltorhinal system in behaving rats.
F1000 Research. 3:98. [PubMed] - *Mizuseki K *Buzsaki G (2013)
Preconfigured, skewed distribution of firing rates in the hippocampus and entorhinal cortex.
Cell Rep. 4, 1010-1021. [PubMed] - *Buzsaki G, Mizuseki K (2014)
The log-dynamic brain: how skewed distributions affect network operations.
Nat. Rev. Neurosci. 15, 264-278. [PubMed] - Belluscio M, Mizuseki K, Schmidt R, Kempter R, *Buzsaki G (2012)
Cross-frequency phase-phase coupling between theta and gamma oscillations in the hippocampus.
J. Neurosci. 32, 423-435. [PubMed] - Schomburg EW, Fernández-Ruiz A, Mizuseki K, Berényi A, Anastassiou CA, Koch C, *Buzsáki G (2014)
Theta phase segregation of input-specific gamma patterns in entorhinal-hippocampal networks.
Neuron 84, 470-485. [PubMed] - Matsumoto N, *Kitanishi T, Mizuseki K (2019) The subiculum: Unique hippocampal hub and more. Neurosci Res 143:1-12. [PubMed]
- Grosmark AD, Mizuseki K, Pastalkova E, Diba K, *Buzsaki G (2012)
REM sleep reorganizes hippocampal excitability.
Neuron 75, 1001-1007. [PubMed] - Sullivan D, Csicsvari J, Mizuseki K, Montgomery S, Diba K, *Buzsaki G (2011)
Relationships between hippocampal sharp waves, ripples, and fast gamma oscillation: influence of dentate and entorhinal cortical activity.
J. Neurosci. 31, 8605-8616. [PubMed] - *Mizuseki K, Miyawaki H (2017) Hippocampal information processing across sleep/wake cycles. Neurosci Res 118:30-47. [PubMed]
- *Mizuseki K, Miyawaki H (2019) Hippocampal information processing and homeostatic regulation during REM and non-REM sleep. In: Handbook of Behavioral Neuroscience 30, pp 49-62: Elsevier. [PubMed]
