Abstract: | ![]() Prior studies of the neural representation of episodic memory in the human hippocampus have identified generic memory signals representing the categorical status of test items (novel vs. repeated), whereas other studies have identified item specific memory signals representing individual test items. Here, we report that both kinds of memory signals can be detected in hippocampal neurons in the same experiment. We recorded single-unit activity from four brain regions (hippocampus, amygdala, anterior cingulate, and prefrontal cortex) of epilepsy patients as they completed a continuous recognition task. The generic signal was found in all four brain regions, whereas the item-specific memory signal was detected only in the hippocampus and reflected sparse coding. That is, for the item-specific signal, each hippocampal neuron responded strongly to a small fraction of repeated words, and each repeated word elicited strong responding in a small fraction of neurons. The neural code was sparse, pattern-separated, and limited to the hippocampus, consistent with longstanding computational models. We suggest that the item-specific episodic memory signal in the hippocampus is fundamental, whereas the more widespread generic memory signal is derivative and is likely used by different areas of the brain to perform memory-related functions that do not require item-specific information.The hippocampus is essential for the formation of declarative (conscious) memory (1, 2), including both episodic memory (memory for events) and semantic memory (factual knowledge). Episodic memories represent the “what, when, and where” information about remembered events (3). Here, we focus on the neural representation of episodic memory for events, specifically words presented and later repeated in a continuous recognition memory format (4).Bilateral hippocampal lesions result in substantial anterograde amnesia for new events, whether memory is tested by recall or recognition (5). By contrast, bilateral lesions to a more anterior medial temporal lobe structure―the amygdala―have no such effect (6). One might therefore expect to find single-unit activity associated with episodic memory in the hippocampus but not in the amygdala. Yet, the earliest single-neuron studies failed to detect hippocampal neurons that fired differentially to recently presented test items vs. novel items. This was true in studies with humans (7, 8) and monkeys (9–11). One early study with monkeys identified a few such neurons in the hippocampus (12), and other studies found them in areas other than the hippocampus (e.g., inferomedial temporal cortex or inferotemporal temporal cortex) (9–11, 13, 14). Overall, this was not the pattern anticipated from lesion studies.Subsequent studies successfully detected some memory-related neural activity (15–17), observing that ∼10% of hippocampal neurons exhibited differential firing rates based on item status, with some firing more for repeated items and others firing more for novel items. Surprisingly, similar “memory-selective” neurons were also reliably detected in the amygdala at approximately the same frequency. Yet, these memory-selective neurons responded differentially to the generic, categorical status of test items (repeated vs. novel) and thus are not episodic memory signals (i.e., signals representing memory for specific events). According to neurocomputational models dating back to Marr (18), episodic memory representations in the hippocampus are supported by sparse neural codes (19–21). If memories for individual items are sparsely coded in largely nonoverlapping (pattern-separated) neural assemblies, it should be possible to find neurons that respond to particular repeated items, rather than to an item’s generic status. Two recent single-unit studies with humans detected such neurons in the hippocampus, but not in the amygdala (22, 23), apparently reflecting sparsely coded episodic memories. In the present study, we tested 1) whether the generic and the item-specific signals coexist in neural firing patterns recorded during the same memory task, and 2) whether the two kinds of signals are present exclusively in the hippocampus or are also evident in other brain regions.During a continuous recognition memory procedure, neurons were simultaneously recorded from four brain regions: hippocampus, amygdala, anterior cingulate cortex, and prefrontal cortex. Altogether, 55 continuous recognition memory sessions were completed by 34 epilepsy patients who had implanted clinical depth electrodes with microwires measuring single-unit activity (SUA) and multiunit activity bilaterally (24). We limited the present analyses to SUA. Words were presented consecutively and repeated once after varying lags; patients judged each word as either “novel” or “repeated.” Thus, repeated words differed from their earlier presentations as novel words only with respect to their combined “what, when, and where” episodic status (3). |