How fast do learning-induced anatomical changes occur in the brain? The traditional view postulates that neocortical memory representations reflect reinstatement processes initiated by the hippocampus and that a genuine physical trace develops only through reactivation over extended periods. Brodt et al. combined functional magnetic resonance imaging (MRI) with diffusion-weighted MRI during an associative declarative learning task to examine experience-dependent structural brain plasticity in human subjects (see the Perspective by Assaf). This plasticity was rapidly induced after learning, persisted for more than 12 hours, drove behavior, and was localized in areas displaying memory-related functional brain activity. These plastic changes in the posterior parietal cortex, and their fast temporal dynamics, challenge traditional views of systems memory consolidation.

Science, this issue p. 1045; see also p. 994



원문: 여기를 클릭하세요~




The representation of memory in the brain is one of the unresolved questions in neuroscience. A key feature of learning and memory is the process of neuroplasticity—the ability of the brain to remodel structurally and functionally as a result of cognitive experience. Although the neurobiological basis of this process (that is, synaptic plasticity) is well established, the system level dynamics of neuroplasticity are still unclear. Recently, diffusion-weighted magnetic resonance imaging (DW-MRI), which can be carried out noninvasively in humans, provided a new approach to explore neuroplasticity. One of the parameters extracted from DW-MRI is mean diffusivity (MD) of water molecules, which is a biomarker of tissue microstructure (1). By probing changes in MD across the brain, areas of microstructural plasticity can be detected within hours of learning. On page 1045 of this issue, Brodt et al. (2) used DW-MRI after an object-location memory task and found that brain plasticity occurs in the posterior parietal cortex (PPC) rather than in the hippocampus. Furthermore, they demonstrated that localized DW-MRI changes follow the basic definition of a memory engram (the physical manifestation of memories stored in the brain). These observations go beyond the understanding of memory representation in the brain. Probing MD with DW-MRI appears to provide a detailed and comprehensive monitoring of brain microanatomy, revealing uncharted relations between brain structure and function.

Memory engrams in the brain

A memory circuit occurs in the hippocampal formation between the parahippocampus (PHC), entorhinal cortex (ENT), and perirhinal cortex (PRC). However, DW-MRI revealed that lasting microstructural changes occurred rapidly in the neocortex as a result of learning and memory tasks, suggesting that memory engrams may exist outside the hippocampus.


Through the evolution of neuroscience, regardless of the methods used, brain mapping theories to explain how the brain functions have led to the development of two camps: localizationism and globalism. Those two camps differ in their approach regarding how functions are mapped in the brain. Whereas localizationists claim that brain functions are localized in specific anatomic regions, globalists suggest that the brain works as a comprehensive unit. The question of localizationism versus globalism has been a long-standing topic of debate in neuropsychology.

Although evidence for functional localization exists for several cognitive domains, the definition of how memory is mapped in the brain remained elusive. The search for the memory engram has been the aim of many researchers. In 1950, Lashley stated that memory is not stored locally but rather spread throughout the cortex (3). Yet, memory disorder studies, such as the studies on patient H.M. who suffered from amnesia after brain surgery to remove the hippocampus, have indicated that the hippocampus is central to memory perception. Using electrophysiology and microscopy, the role of the hippocampus in different types of memory representations became axiomatic. A detailed map of interactions between the different parts of the hippocampus and different phases of memory was created (see the figure). Consequently, the neocortex was considered to imprint memories more slowly than the main memory hippocampal hub. Functional MRI (fMRI), which noninvasively measures brain function by monitoring blood flow and volume changes, followed the path of localizationism and has revolutionized human brain mapping and provided an invaluable noninvasive tool to explore task-specific brain activity. fMRI provides a more detailed and empiric map of brain function than was previously attainable. Such studies supported the concept of the hippocampus as the potential center of the memory engram.

Yet, the study of Brodt et al. gives solid experimental proof that a memory engram is present in the PPC, at least with respect to the object-location memory task used (2). This suggests that the undisputed role of the hippocampus in learning and memory may be in doubt. Why has this striking observation not been made before? Brodt et al. used a different approach to explore the memory engram—characterizing memory-related microstructural plasticity through MD measurements from DW-MRI (45). MD is a nonspecific, but sensitive, marker of tissue microstructure and has been used primarily as an indicator of inflammation and less frequently in conventional brain mapping. Studies in rats have suggested that the neurobiological basis of the rapid microstructural changes that occur after experience-driven neuroplasticity that is measured by MD may be related to astrocyte (a type of glial cell) remodeling (45). Although MD cannot be as specific as advanced microscopy methods, it provides unprecedented high sensitivity to in vivo system-level neuroplasticity in humans.

Recent studies have explored the utility of MD as a rapid and sensitive marker of neuroplasticity in several types of learning: spatial navigation (5), motor sequence learning (6), phonological language learning (7), and object-location memory (2). In these studies, the tasks lasted only hours, and, surprisingly, it appears that hippocampal involvement is less compared to other regions. Yet, as noted by Brodt et al., it is not that the hippocampus is not involved but rather that the changes in the hippocampus are more frequent and diminish rapidly. Indeed, MD measurements in rats undertaking a rapid spatial navigation task revealed that after a few episodes of training, most MD changes are observed in the hippocampus, but additional training, within the time frame of short-term memory, shifts most of the changes to the neocortex (8).

The use of MD and memory tasks revealed memory-associated changes in various neocortical regions (which may be newly identified memory engrams). Although these changes were previously shown in rodents (9), MD enables the detection of such engrams in humans as well. Moreover, coupled with fMRI, MD measurements from DW-MRI can detect the system-level dynamics of memory formation and storage, indicating that the transfer of memory from the hippocampus to the cortex occurs within minutes, much faster than previously thought. That MD can be used to detect such phenomena underscores its utility: It is a microstructural probe measured on the macroscale.

The ability of MD to detect microstructural imprinting in various neocortical regions exemplifies the localization versus globalization views of brain mapping. Without disproving previous theories and observations, the studies of Brodt et al. and others who have used MD postulate that brain mapping is an issue of scales and dimensions. Memory representation can be viewed as a task-specific regional process or as a global feature of the brain that can be segmented into localized domains.

In view of the ability of MD to detect time-dependent, system-level tissue changes, we anticipate that MD will highlight memory processes that have not yet been charted in the human brain. These processes may reveal new features of brain function-structure-cognition relationships that will improve our understanding of the localizationism-globalism symbiosis in brain architecture, network, and hierarchy and enable us to clarify with greater accuracy where memory as well as other cognitive domains are represented in the brain.



원문: 여기를 클릭하세요~

Leave a Reply

Your email address will not be published. Required fields are marked *