The ability to recognize individuals, retrieve associated value information, and update this information is crucial for forming social relationships. Animals constantly interact within social groups, recognizing individuals, recalling past interactions, and incorporating new information. Associating individuals with subjective value and updating that value is key to social relationship formation. The hippocampus is known for processing episodic information, including the 'where, when, and what' of experiences, and its role in social information processing is increasingly recognized. Several studies have shown the hippocampus's importance in social memory, and this study focuses on understanding how the hippocampus encodes both social identity and reward value to support associative social memory. Prior research has shown that the hippocampus, specifically the CA1 region, plays a significant role in processing reward-related information, particularly in the context of spatial navigation. However, its role in associating reward with social information is less understood. Evidence suggests a link between the hippocampus and associating reward with social information, particularly in paradigms like social transmission of food preference and in studies involving familiar conspecifics with positive or negative associations. Given the evidence of reward processing in the dorsal CA1 and its projections from the socially-relevant dorsal CA2, the dorsal CA1 is a strong candidate for integrating social identity and reward value. This study aimed to investigate this hypothesis by developing novel Go-NoGo paradigms that require mice to associate stimulus mice with reward availability.

Existing literature extensively documents the hippocampus's role in episodic memory, encompassing spatial and temporal aspects of experiences. The hippocampus's involvement in social memory is also supported by studies demonstrating its importance in recognizing familiar individuals and associating them with positive or negative experiences. Studies using social transmission of food preference paradigms have shown the hippocampus's importance in associating food scents with social cues. Similarly, research with golden hamsters has shown dorsal CA1's role in associating social identities with negative values. Human studies further highlight the hippocampus's role in processing social affiliation and power dynamics. This study builds on these findings by focusing on the integration of social identity and reward value within the dorsal CA1 region, a key area for both social and reward processing.

The researchers developed novel Go-NoGo individual discrimination paradigms. Male subject mice were trained to discriminate between two familiar male littermate mice based on their individual characteristics, not broader social categories. One stimulus mouse was associated with a water reward (Go trial), and the other was not (NoGo trial). Mice first underwent rule learning, discriminating between a mouse and an empty head-fixing device. Successful mice then proceeded to the individual discrimination phase, reaching a criterion of >80% accuracy. To assess the role of visual cues, mice were tested under alternating light and dark conditions. To assess long-term memory, mice were tested after a 72-hour break. The effect of dorsal CA1 inactivation was investigated via bilateral muscimol injections. Two-photon calcium imaging in Thy1-GCaMP6f transgenic mice monitored dorsal CA1 neuronal activity during the task, including reversal learning (changing reward associations). The researchers examined neuronal activity patterns associated with Go and NoGo trials and across reversal learning sessions. To distinguish between stimulus-driven and reward-expectation-driven neuronal activity, sessions with reversed reward contingencies were compared. A support vector machine (SVM) was used to decode trial types (Go/NoGo) from neuronal activity. To investigate whether reward expectation was specific to social tasks, a similar non-social odor discrimination task was conducted. Finally, a four-mouse discrimination task was developed to better identify individual mouse-specific neural activity, employing four stimulus mice, two associated with rewards and two not. This allowed for the identification of individual-specific neural responses independent of reward prediction. This four-mouse task also investigated long-term memory (6-day break) and the effects of dorsal CA1 inactivation. Two-photon calcium imaging in the four-mouse paradigm was used to examine individual mouse discrimination at both single-neuron and population levels. The stability of individual mouse-specific neural activity was assessed across multiple days. Similar analyses were performed for non-social odor discrimination tasks.

The study found that mice could discriminate between individual conspecifics based solely on olfactory cues from a brief nose-to-nose investigation (approximately 0.5 seconds), and this ability was dependent on the dorsal CA1 hippocampus. Inactivation of the dorsal CA1 significantly impaired performance in the individual discrimination task but not in a non-social odor discrimination task. Two-photon calcium imaging revealed that a subset of dorsal CA1 neurons reliably represented reward expectation, regardless of the specific mouse associated with the reward, and this representation persisted over several days. This suggests that the reward-associated activity is not tied to specific social identities but rather to the anticipated reward itself. Furthermore, a dynamic subset of CA1 neurons showed high accuracy in discriminating between individual mice. This population of neurons showed a high degree of plasticity, with the specific neurons representing each individual changing from day to day, suggesting a flexible representation of social identities. In contrast to the consistent reward representation, this dynamic representation suggests that the social identity information is not persistently encoded in the dorsal CA1. In the four-mouse discrimination task, which did not involve reversal learning, individual mouse discrimination was observed in dorsal CA1 neurons, both for reward-associated and non-reward-associated mice. Furthermore, population activity successfully decoded individual mouse identity with high accuracy. Even without a task, passive viewing of novel mice elicited individual-specific responses in some CA1 neurons, although at a lower rate than during task engagement. Stability analysis across days revealed that the neural representation of individual mice was dynamic, with some level of long-term stability at both single-cell and population levels. This dynamic representation of social identity contrasts sharply with the stable representation of reward expectation. Analyses of non-social odor discrimination tasks showed that CA1 neurons also represented non-social odors, but the representation of reward expectation was absent in non-social tasks. Comparing social and non-social tasks, the stability of neuronal representations was significantly higher for non-social odors than social identities. This suggests that the dorsal CA1 plays a role in forming associations between social identities and reward values but uses different mechanisms for representing social identities compared to non-social stimuli.

The findings demonstrate that the dorsal CA1 hippocampus plays a crucial role in associative social memory by integrating social identity and reward expectation. The stable representation of reward expectation suggests a dedicated neural population for encoding reward prediction, independent of the specific social context. The dynamic representation of social identity suggests a flexible coding mechanism, possibly relying on information from other brain regions such as the ventral CA1, dorsal CA2, and medial prefrontal cortex. The absence of reward-selective neurons in the non-social odor discrimination task highlights the specificity of this reward encoding to social interactions. The study's methodology addresses potential confounds by controlling for visual cues and using a four-mouse task that allows for individual discrimination independent of reward reinforcement. This study expands our understanding of hippocampal function beyond spatial memory, highlighting its role in complex social cognition. The findings underscore the hippocampus's flexibility in encoding both stable and dynamic representations, accommodating the need for both consistent reward expectations and adaptable social information processing.

This study provides compelling evidence for the dorsal CA1 hippocampus's involvement in associative social memory, showing distinct neural representations for reward expectation (stable) and social identity (dynamic). The dynamic nature of social identity encoding suggests a flexible system that integrates information from other brain areas. Future research could investigate the interaction between the dorsal CA1 and other brain regions involved in social memory, as well as explore the underlying mechanisms driving the dynamic representation of social identities.

The study primarily used male mice, limiting the generalizability to females. The use of head-fixed mice, while controlling for extraneous factors, might not fully replicate natural social interactions. The specific olfactory cues driving individual recognition remain to be fully elucidated. Furthermore, the study focused on associative social memory within a specific task paradigm, and the findings may not generalize to all forms of social memory.