From necessity to sufficiency in memory research: when sleep helps to understand wake experiences Marie Masako Lacroix * , Gaetan De Laville ´ on * and Karim Benchenane Memory is the ability to adapt our behavior by using the stored information, previously encoded. The first investigations of the neuronal bases of the memory trace concerned its properties (location, cellular and molecular mechanisms, among others). However, to understand how this is achieved at the scale of neurons, we must provide evidence about the necessity of a neuronal subpopulation to support the memory trace, but also its sufficiency. Here, we will present past and recent studies that provide information about the neuronal nature of memories. We will show that research on sleep, when cells assembly supposedly carrying information from the past are replayed, could also provide valuable information about the memory processes at stake during wake. Address Team Memory, Oscillations and Brain states (MOBs), Brain Plasticity Unit, CNRS UMR 8249, Ecole Supe ´ rieure de Physique et de Chimie Industrielles de la Ville de Paris - ParisTech, Paris, France Corresponding author: Benchenane, Karim (karim.benchenane@espci.fr, benchenane.karim@gmail.com) * Both authors contributed equally. Current Opinion in Neurobiology 2015, 35:156–162 This review comes from a themed issue on Circuit plasticity and memory Edited by Thomas Mrsic-Flogel and Alessandro Treves http://dx.doi.org/10.1016/j.conb.2015.08.009 0959-4388/# 2015 Elsevier Ltd. All rights reserved. Memory is supported by the experience-induced mod- ifications in our brain that can be accessed later and used to adapt our behavior. Semon first proposed that memory was supported by an ‘engram’ (cited in [1]), correspond- ing to the dynamic changes made to the brain that enable the storage of one memory. Hebb further proposed that modification of the synaptic connection between neurons supports memory formation and that the engram relies on the ensemble of mono-synaptically connected neurons (called cell assembly) associated with an experience [2]. As mentioned by Yadin Dudai, the term engram is often used interchangeably with memory trace [3]. Declarative/explicit versus non-declarative/ implicit memory The seminal study of patient HM provided evidence for the existence of several forms of memory: declarative/ explicit knowledge relies on hippocampus integrity while non-declarative/implicit one does not [4,5]. Declarative/ explicit memory has been defined as a capacity to verbally report specific past experiences or factual knowledge, corresponding respectively to episodic and semantic memory. In both cases, a conscious explicit recall is needed to be qualified as declarative. By contrast, non- declarative or implicit memory does not necessitate these conscious processes. In order to study declarative-like memory in rodents, researchers have developed different tasks during which rodents’ behavior is supposed to be indicative of their knowledge about the world or their past experience. Importantly, the task must not be solvable by using automatic behavior relying on non-declarative memory and should be solved by a goal-directed strategy. Accord- ing to Dickinson and Balleine, an action is goal-directed if its performance is controlled by the representation of both the contingency between the action and the outcome, and the representation of the outcome as a goal [6]. The standard way of determining the goal dependency of an action is to conduct an outcome devaluation test or a modification of the action–outcome contingency. More- over, animal models of declarative memory should be, as in human, sensitive to hippocampal lesions while non- declarative memory should be spared. The two most widely used tasks in memory research are the Morris water maze in which rodents must find an invisible platform by using external cues and the contex- tual fear conditioning where foot-shocks are delivered in a given spatial context. Importantly, a double dissociation similar to what observed in humans has been observed in rodents [7,8]. Based on these tasks (and others) it was shown that the hippocampus is involved in spatial mem- ory relying on the use of external cues (allocentric). O’Keefe and Nadel proposed the concept of the hippo- campus-dependent cognitive map after the discovery of place cells, hippocampal neurons that fire when the animal is in a given part of an environment [9]. Yet, a place cell could fire in different environments (a phenom- enon called remapping) and the precise localization in different environments is encoded by a given correlation pattern at the population level [10]. Available online at www.sciencedirect.com ScienceDirect Current Opinion in Neurobiology 2015, 35:156–162 www.sciencedirect.com