The hippocampal formation is a brain region that plays critical roles in memory and spatial navigation. A mechanistic understanding of hippocampal circuit organization and function is fundamental for determining how this brain region contributes to memory and cognition. Although the general anatomy and circuit organization of the hippocampal CA1 has been well-studied, most of our understanding of hippocampal circuits comes from the conventional anatomical tracing studies which lack cell-type specificity and quantitative measurements of connectional strengths. New advances in virology and genetics complement traditional approaches and are powerful tools for mapping cell-type-specific circuit connectivity and function. Herein, through a series of extensive studies using cutting-edge viral and genetic techniques, novel hippocampal circuits and their functions are elucidated. In Chapter 1, a Cre-dependent, genetically modified rabies-based tracing system was developed to map local and long-range monosynaptic connections to specific excitatory and inhibitory CA1 neuron types in the mouse. Our data show the different input sources of varying strengths are distributed onto each specific CA1 cell type, providing insight into differential circuit mechanisms of hippocampal functional operations. In Chapter 2, quantitative re-evaluation of intra- and para- hipppocampal input connections to excitatory neurons in different CA1 proximodistal subfields was performed using monosynaptic rabies tracing. The results provide a new topographic circuit basis for functional considerations of CA1-associated memory and cognition. For the studies reported in Chapters 3 and 4, a detailed analysis of synaptic circuit organization and function of the subiculum to CA1 back-projection pathway was performed using state-of-the-art techniques including combinatorial viral tracing, genetically targeted manipulation of neural activity, and behavioral analysis. Building on the results from chapter 1 that provide unambiguous anatomical evidence for non-canonical subicular back-projections to CA1, global circuit input and output connections of CA1-projecting and other subicular neurons were mapped and compared. These studies establish that CA1-projecting subicular neurons are a distinct neuronal group with unique circuit properties within the subiculum. To link circuit mapping to function and behavior, I investigated how DREADDs-mediated inactivation of CA1-projecting subicular neurons modulates spatial learning and memory, and found that inactivating CA1-projecting subicular neurons specifically impairs animal’s object location memory. This study has, for the first time, implicated the non-canonical subicular projections in hippocampus-associated spatial memory behavior. Together, this dissertation has provided novel, cell-type-specific anatomical and functional insights for hippocampal CA1 and the subiculum, and addressed the circuit organization and function of the under-appreciated bidirectional connections of subiculum and hippocampal CA1. This study may also lead to a better understanding of the neural circuit mechanisms that underlie hippocampal-related neurological disorders such as Alzheimer's disease.

The hippocampal CA1 subregion is critical for processing spatial and episodic memories. As the principal cell type, CA1 pyramidal neurons have been found to represent variables that are essential for spatial navigation and elements in episodic memory. At the same time, multiple CA1 pyramidal neurons can display synchronized activations, which can represent complex memory elements that individual neurons cannot process. An important question to date is how the co-activated CA1 pyramidal neurons are spatially organized in anatomical space. Existing research on this question had provided contradictory results, and were all subjected to limitations, for example, the limited anatomical resolution for the electrophysiological-based studies, and the limited behavior for the two-photon based studies. Another question is how the organized co-activations of CA1 pyramidal neurons are affected by Alzheimer’s disease (AD). Existing studies have revealed that AD can induce aberrant activation pattern in individual pyramidal neurons, and has negative impact on the circuit connectivity, but no study have investigated AD’s potential influence on the anatomical organizations of the co-activating neuron subgroups. Here I utilized a novel imaging technique, the head-mounted miniature microscope (Miniscope), to investigate the ensemble activities of CA1 pyramidal neurons with the limitations above largely addressed. In Chapter 1, I used Miniscope to record neural ensemble activities in mouse hippocampal CA1 and identified sub-populations of excitatory neurons that co-active across the same second-long period of time while also clustered in anatomical space. This kind of organization vary in membership and activity dynamics with respect to movement in different environments, and appear during immobility in the dark, suggesting internal mechanisms that guide their formation. In Chapter 2, I used Miniscope to investigate the populational CA1 neural activities in both wildtype and 3xTg AD mice, an Alzheimer’s disease model mouse type contains three mutations associated with familial Alzheimer's disease. I identified hyperactivations among the CA1 pyramidal neuron populations in AD mice at different ages in open field exploration, as well as impaired spatial representation illustrated by lower spatial information score and higher sparsity. Having established Miniscope’s ability to identify anatomical organization of co-activated CA1 pyramidal neurons, and to reveal the difference of their firing profile between control and AD genotypes, I examined the discovered temporal-anatomical pyramidal neuron clusters to both wildtype (WT) and 5xFAD mice under different ages. I found the clusters existed in both genotypes and no difference was noted for intra-cluster pairwise correlation and anatomical cluster size. When examining the cluster-level ensemble representation of the environment, young AD mice show a lower level of specificity between the labelling of different clusters, but no differences were noted between WT and AD for other age groups. Both WT and AD displayed elevated cluster dissimilarity between linear tracks with 90-degree direction difference, but only at 8~10-month age. Finally, the segregation level of the anatomical clusters, measured by the mosaic level metric, displayed significant negative correlation with the averaged pairwise correlation strength in young and mid WT mice, as well as young AD mice, but showed no significant relationship with the correlation of spatial rate maps. Together, the thesis describes a previously under studied organization of CA1 pyramidal neurons and performs preliminary investigation of how this organization behaves under AD conditions. The results should contribute to our understanding of CA1 neural organizations that support the region’s memory processing function.

The hippocampal formation (HF) and the retrosplenial cortex (RSC) are complex brain regions involved in learning, memory, and cognitive functions, such as spatial navigation and planning. While both regions exhibit diverse circuit connectivity, the organization of these circuits and their impact on behavior remain unclear. To address these knowledge gaps, I have applied viral-genetic mapping techniques to investigating the connectivity of the HF and RSC and their associated behavioral functions. Recent previous studies of the HF identified noncanonical pathways involving the subicular complex and hippocampal subregions CA1, suggesting the existence of extended noncanonical networks involving additional brain regions. For the study described in Chapter 1, I used multiple retrograde and anterograde viral tracers to quantitatively map the circuit connectivity of excitatory neurons in the dorsal hippocampal CA3 subregions. My results reveal significant noncanonical synaptic inputs from ventral CA1, perirhinal cortex, and the subicular complex to dorsal CA3 that follow a proximodistal topographic gradient with regard to CA3 subregions. I confirmed the functional contributions of these pathways: genetic inactivation of the ventral CA1 to dorsal CA3 projection impairs object-related spatial learning and memory. For the study described in Chapter 2, I employed monosynaptic rabies virus tracing to conduct a quantitative examination of the local and global inputs to CA3 inhibitory neurons. Similar to the results of earlier studies of CA3 excitatory tracing, my findings indicate that CA3 inhibitory neurons also receive noncanonical circuit inputs from ventral CA1 and the subiculum complex. Interestingly, my detailed study reveals that these inputs exhibit a topographical gradient along the proximal/distal axes. My study's results provide a new anatomical connectivity basis for studying the function of CA3 excitatory and inhibitory neurons, as well as a circuit foundation to explore their novel roles in hippocampal circuit dynamics and learning and memory behaviors. For my study described in Chapter 3, I employed retrograde and anterograde viral tracers in conjunction with monosynaptic retrograde rabies virus to investigate differential connectivity and functionality among distinct subpopulations of RSC neurons based on their projection patterns. My results reveal notable differences between the secondary motor cortex (M2)-projecting RSC neurons and anterior dorsal thalamus (AD)-projecting RSC neurons. Specifically, M2-projecting neurons exhibit more extensive afferent input from the dorsal subiculum, lateral dorsal and lateral posterior thalamus, as well as sensory cortices, as compared to AD-projecting neurons. Chemogenetic inhibition of M2-projecting RSC neurons leads to impairments in both object location memory and place-action association, whereas the RSC to AD pathway specifically affects object-location memory. These results demonstrate that RSC connectivity and function are organized as a collection of semi-independent circuits. My functional studies show that these semi-independent RSC circuits integrate information from distinct sets of RSC afferents. Together, by elucidating the circuit organization and functional roles of the HF and RSC, my studies shed light on the complex cortico-hippocampal circuit connectivity organization underlying learning and navigation.

Developing neuronal circuits have many interesting properties; understanding the development process can give us insights on functioning of complex adult neural circuits. One of major features of immature neural circuits is the spontaneous activity, which has also been considered a driving force of circuit maturation. Giant Depolarizing Current (GDP) is the major spontaneous activity at single neuron level of developing hippocampus. What is missing, however, is whether hippocampus has large scale network level spontaneous activity that precedes the formation of mature hippocampal circuits, i.e. unidirectional trisynaptic flow. In chapter 1 of this study, we use Voltage Sensitive Dye Imaging to investigate the initiation and propagation of global network activity (GNA) in developing hippocampus. GNA originates in distal CA3 and propagates both forward to CA1 and backward to DG. Single-cell and local field potential recording confirmed GNA is closely related to neuronal activation. Further, enhancement of local circuit connections to excitatory pyramidal neurons occurs over the same time course as GNA. Thus Bi-directional GNA precedes the maturation of the mouse hippocampal circuit, and it is correlated to maturation of functional circuit connections. To further understand the underlying pathway for back-projection from CA3 to DG in the developing hippocampus, in the study of chapter 2, we use laser scanning photostimulation technique to map synaptic circuit development of mossy cells. We observed major regional sources of circuit excitation and inhibition to hilar mossy cells were changed from CA3 to DG during development. The excitatory back projection from CA3 to mossy cell were greatly reduced in adult hippocampus. Thus, hilar mossy cells can play important role in circuit signal back-propagation from CA3 to DG. In chapters 3 and 4, we demonstrate an innovative technology which automatically detects and extracts EPSC and IPSC events in electrophysiological recording. We also present our GPU based computing system and our modification of the previous technology to utilize the power of GPU computing. GPU computation could greatly improve the speed of computation while retaining data precision. Taken together, my dissertation studies contribute new and important knowledge to the field of hippocampus circuit studies, using our newly developed techniques.

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder with a growingmajor health impact, particularly in countries with aging populations. Traditional therapeutic approaches have been taken towards treating AD by focusing on neurochemical and neuropathological mechanisms. The examination of neural circuit mechanisms in AD mouse models is a recent focus for identifying new AD treatment strategies. Accumulating evidence shows that there are neural circuit-level maladaptive alterations in AD brains. New viral-genetic technologies facilitate quantitative mapping of cell-type-specific neural circuit connections in AD mouse models. Monosynaptic rabies virus mapping reveals age-progressive changes in both long-range and local hippocampal neural circuit connections that occur in AD mouse models – and provides neural circuit-based explanations for human AD behavioral defects. The recent developments in concepts and technology present new opportunities for studying AD pathogenesis at the neural circuit level. In Chapter 1, I provide a literature overview of new technical advancements in neural circuit mapping, describe both conventional neural tracers and state-of-the-art virus-based tracers, and discuss relevant applications using the novel viral- genetic tracers. In Chapter 2, I report our investigation of neural circuit connectivity alterations of hippocampal CA1 excitatory neurons in AD model mice. I focus on the hippocampus as it is one of the most affected brain regions in AD. I applied genetically targeted monosynaptic rabies viral xii tracing to map the neural circuit connectivity of CA1 excitatory neurons in age-matched APP-KI model mice versus wild-type (WT) control mice. Our quantitative analysis of neural circuit connections reveals that compared to WT mice, APP-KI mice show age-progressive and sex- dependent hippocampal CA1 connectivity alterations. Unexpectedly, AD mice exhibit CA1 excitatory neuron input shift with hippocampal CA3 in AD mice. Moreover, aged female AD mice show significantly reduced subiculum to CA1 projections. This study is the first to characterize hippocampal CA1 neural circuit changes in AD model mice using monosynaptic rabies viral tracing. In Chapter 3, I report our study of using monosynaptic rabies virus to map the neural circuit connectivity of the subiculum excitatory neurons in age-matched, gender-balanced control and 5xFAD mice. The subiculum is a critical brain region for relaying and integrating hippocampal and cortical information, and is among the most vulnerable brain regions in AD. My study shows that the neural connectivity of the subiculum exhibits a significant sex- and age-dependent differential pattern in the 5xFAD mice. I find an unexpected increase in visual cortex input connectivity to the subiculum in aged 5xFAD mice. Our new findings are supported by human AD literature and can help to identify potential new therapeutic circuit targets for AD treatments. Overall, the studies presented in this dissertation provide new insights into neural circuit mechanisms underlying AD pathogenesis and support the notion that neural circuit disruptions are a prominent feature of AD.

Background

New method

Results

Comparison with existing method(s)

Conclusions

Background

New method

Results

Comparison with existing method(s)

Conclusions