A Gradient of Hippocampal Inputs to the Medial Mesocortex

Abstract

Running title: Hippocampal inputs to the medial mesocortex Key-wordsHippocampus, anterior cingulate cortex (CG), retrosplenial cortex (RSC), medial mesocortex (MMC), hippocampal inputs, optogenetics, in vitro and awake-behaving multi-electrode recordings, long-range inhibitory neurons HighlightsBoth MMC regions CG and RSC receive monosynaptic connections from the dorsalintermediate CA1 CG receives layer-sparse excitatory projections exclusively originated from stratum piramidale whereas RSC is targeted densely in superficial layers by a mixed excitatory and inhibitory input originating from all CA1 strata CA1 monosynaptic projections correspond to active synapses onto distinct layers of the two MMC regions Diverse synchrony between MMC and HIPP recorded in vivo is consistent with the rostro-caudal diversity of direct HIPP-MMC connections AbstractMemory-guided decisions depend on complex, finely tuned interactions between hippocampus and medial mesocortical regions anterior cingulate and retrosplenial 2 cortices. The functional circuitry underlying these interactions is unclear. Using viral anatomical tracing, in vitro and in vivo electrophysiology, and optogenetics, we show that such circuitry is characterized by a functional-anatomical gradient. While CG receives excitatory projections from dorsal-intermediate CA1 originated exclusively in stratum pyramidale, retrosplenial cortex also receives inputs originating in stratum radiatum and lacunosum-moleculare, including GAD+ neurons providing long-range GABAergic projections. Such hippocampal projections establish bona fide synapses throughout cortical layers, with retrosplenial cortex densely targeted on its layer 3, around which it receives a combination of inhibitory and excitatory synapses. This gradient is reflected in the pattern of spontaneous oscillatory synchronicity found in the awake-behaving animal, compatible with the known functional similarity of hippocampus with retrosplenial cortex, which contrasts with the encoding of actions and "task-space" by cingulate cortex. Katche et al., 2013; Vann and Aggleton, 2005), and the former necessary for flexible choice behaviors and long-term memory (Chudasama et al., 2003;Goshen et al., 2011;Maviel et al., 2004; Teixeira et al., 2006; Vetere et al., 2011; Wang et al., 2012).These findings agree with the few studies recording MMC neural activity in vivo, and with the patterns of HIPP-MMC functional coordination Wilson, 2013, 2015). Neurons in CG encode choice effort (Cowen et al., 2012;Endepols et al., 2010) and task-space by assimilating into motor decisions the spatial-contextual properties of available choices (Remondes and Wilson, 2013;Sul et al., 2010), while RSC joins visual-spatial information such as head-direction and movement in space (Chen et al., 1994a(Chen et al., , 1994bCho and Sharp, 2001) to encode spatial landmarks (Alexander and Nitz, 2015;Auger et al., 2012;Mao et al., 2017). The above findings are supported by the patterns of functional interaction between MMC and HIPP in vivo, namely the d...