Invertebrate Learning and Memory

Memories exist in multiple forms and have multiple functions. They may be categorized according to their cell-physiological substrates along a timescale defined as short-term, mid-term, and long-term memory (STM, MTM, and LTM, respectively). Ongoing neural activity serves as the storage device for STM; intracellular signaling cascades lead to MTM; and gene activation, protein synthesis, and new structures underlie LTM. The physiological substrates of these memory stages or phases can be sequential or in parallel, indicating that memory systems are highly dynamic. But what exactly is processed and stored in these different phases of memory formation? Memory is about something, namely objects, events, and relations between objects in the external world as well as internal body states such as hunger and satiety. Thus, memory stores information about the meaning of multiple signals, external and internal; in other words, it has content. Stimuli and actions are evaluated by the nervous system according to expected outcomes, and it is this loop into the future that defines the core of memory content. The ultimate goal of memory research is to uncover the neural mechanisms that allow the content of memory to be encoded, stored, processed, and retrieved. The content of memory is usually considered to be encoded as an engram or memory trace. Lashley referred to the engram in the title of his famous paper and asked questions concerning the location of the engram(s) in different parts of the mammalian brain (cortex). Localization is indeed a major feature of memory, and in the mammalian brain memory localizations can be categorized according to the types of memory that are processed—for example, procedural memory (e.g., cerebellum), episodic memory (hippocampus and prefrontal cortex), and emotional memory (amygdala)—but the content of each of these memories involves many more parts of the brain. Another character of memory is content-sensitive processing. Any retrieval from the memory store changes its content due to the updating process in working memory, a process referred to as ‘reconsolidation.’ It is this updating process that may reveal rules underlying generalization, categorization, and implicit (and explicit in humans) forms of abstraction. However, both localization and content-specific processing tell us little about the mechanisms of how content is encoded and stored in the nervous system, although knowledge of both aspects of the engram is requirements for hypothesisdriven research. Cognitive psychology has struggled with the question of whether the engram or memory trace ‘exists’ if it is not retrieved. “Where is the memory trace when we are not remembering? . . . The hunt for the engram (the physical manifestation of the memory trace that is independent of the operations needed to recover it) may prove to be fruitless as the hunting of the Snark”. Indeed, the engram is not yet a memory if it is not activated, but it is the necessary physical condition for memory to emerge through the readout process in the nervous system. In this sense, the engram, together with the neural processes of activating it and combining it with the information provided by the retrieval process, leads to memory. The informational content of the engram is therefore rather elusive, and we characterize our efforts better by saying that we aim to uncover necessary physical components that we hope will define (at least to some extent) the informational content of the trace leading to the engram. These attempts will be very limited because in reading these physical components as separate entities, the emerging properties from the interaction of multiple components necessary to