Visit the Davachi Lab
How are memories formed? Why do we only remember some of what we encounter? Why do we remember some events in exquisitely rich detail, only have a sense or feeling that we’ve encountered other events and still forget others entirely?
These are some of the questions that we are exploring in the laboratory. Our laboratory complements cognitive behavioral approaches with a pursuit of knowledge of how the brain carries out mnemonic functions (using fMRI) with the belief that this approach will lead to a unified study of memory in which structure (i.e. the brain) and function (i.e. behavior) will finally be linked. This union will allow for more complex hypothesis testing than either approach on its own.
Since the landmark case of patient H.M., long-term memory formation has been intimately linked to the medial temporal lobe. However, despite the fact that the hippocampus is likely the most studied structure in the brain, the specific contributions of individual structures within the medial temporal lobe and how they interact still remain to be well described. Our laboratory is investigating how the medial temporal lobe supports memory formation, the different kinds or forms of memory supported by this system and how this system interacts with other processing systems (e.g. attentional and emotional systems). Below are examples of some of the questions that we (and others) have long been interested in and how we have gone about studying them.
Recognition and recollection
Sometime we know that we have encountered an item in our past (i.e. we recognize it) and other times we remember vivid contextual details surrounding our encounter with that item (i.e. when and where we encountered it, who else may have been around, what we were thinking or feeling at the time…etc). Much research has focused on whether these two memory outcomes can arise from the same underlying process (i.e. one perhaps supported by a single functional memory system) or are, indeed, the consequence of a two distinct processes.
Our approach has been to use brain data to inform this question. It has been hypothesized that the hippocampus and cortical medial temporal regions (e.g. entorhinal, perirhinal and parahippocampal cortices) might support different kinds of episodic learning. In particular, it has been suggested that the perirhinal cortex might provide a substrate for item recognition, while hippocampal processing might provide the substrate for rapid binding of contextual information. We hypothesized that the hippocampus might then be responsible for incorporating contextual information into the memory trace while, perhaps, other cortical regions such as the perirhinal cortex might be responsible for the encoding of item information (in a sense, devoid of context).
We tested this by having subjects encounter information (in this case, a list of words) while being scanned using fMRI. They performed one of two distinct cognitive tasks (e.g. an imagery task or a phonological task) with each item. We reasoned that the cognitive operations being performed with each word would provide different "cognitive" contexts (as opposed to other types of context which might be harder to implement in a scanner environment). Subjects' memory was then tested for both item information (did you see this word?) and context information (which task did you perform with this word?). The data showed a striking distinction between the kinds of information predicted by hippocampal and perirhinal cortical activation at encoding. Specifically, activation of the hippocampus correlated with whether subjects later successfully recalled the context in which they encountered the word while perirhinal activation correlated with whether subjects were going to recognize each word, but not whether they would recall the context. This was the first study in humans to demonstrate a clear distinction between perirhinal and hippocampal encoding processes (Davachi et al., 2003) and suggests that processes leading to item recognition are distinct from those leading to detailed contextual recollection. These data thus provide strong support for dual process theories of recognition.
Working memory and episodic encoding
Working memory has been defined as the short-term maintenance and manipulation of information. For example, you may engage in maintenance while rehearsing your list of groceries in your mind on your way to the grocery store. Imagine you then get a phone call from a friend with whom you are planning on cooking dinner that night and she tells you that she has certain ingredients and is missing others. As you add and subtract from the original list in your mind, you are actively manipulating this information. Both of these processes fall under the rubric of working memory. It has long been debated whether simple maintenance of information (i.e. rehearsal) impacts long-term memory formation. We have shown that the greater the activation of brain regions known to be involved in rehearsal (e.g. prefrontal, parietal, cerebellum), the more likely subjects were to later recognize having encountered that particular item (Davachi et al., 2001). This data suggest that item rehearsal can serve as a conduit into long-term memory.
Our main goals for future research are to (1) further probe the distinct roles that medial temporal lobe regions may play in memory formation, (2) examine whether a functional architecture exists within the hippocampus and what particular kinds of context are supported by this system and (3) investigate the specificity of the medial temporal lobe memory system to long-term declarative memory as there is increasing evidence that it may play a role in short-term or working memory processes as well as implicit learning.
Ph.D. Yale University 1999
Fellowships and Awards
2003 Summer Institute in Cognitive Neuroscience Fellowship, Lake Tahoe
Davachi, L. (2004) The ensemble that plays together, stays together, Hippocampus, 14:1-3.
Davachi, L., Mitchell, J.P. and Wagner, A.D. (2003). Multiple learning mechanisms: distinct medial temporal processes build item and source memories. PNAS, 100(4): 2157-2162
Davachi, L. and Wagner, A.D. (2002). Hippocampal contributions to episodic encoding: insights from relational and item-based learning. Journal of Neurophysiology, 88: 982-990.
Davachi, L., Maril, A. and Wagner, A.D. (2001). When keeping in mind supports later bringing to mind: neural markers of phonological rehearsal predict subsequent remembering. Journal of Cognitive Neuroscience, 13:1059-1070.
Wagner, A.D. and Davachi, L. (2001). Cognitive neuroscience: forgetting of things past. Current Biology, 11: R964-967.
Davachi, L. and Goldman-Rakic, P.S. (2001). Primate rhinal cortex participates in both visual recognition and working memory tasks: functional mapping with 2-DG. Journal of Neurophysiology, 85: 2590-2601.
Sybirska, E., Davachi, L. and Goldman-Rakic, P.S. (2000). Prominence of direct entorhinal-CA1 pathway activation by cognitive tasks revealed by 2-DG functional mapping in the nonhuman primate. Journal of Neuroscience, 20: 5827-5834.
Levy, R., Friedman, H.R., Davachi, L. and Goldman-Rakic, P.S. (1997). Differential activation of the caudate nucleus in primates performing spatial and nonspatial working memory tasks. Journal of Neuroscience, 17: 3870-3882.
Rangarajan, A., Chui, H., Mjolsness, E., Pappu, S., Davachi, L., Goldman-Rakic P.S. and Duncan, J. (1996/7). A robust point-matching algorithm for autoradiographic alignment. Medical Image Analysis, 1: 379-398.
Carden, S.E., Davachi, L. and Hofer, M.A. (1994). U50-488 increases ultrasonic vocalizations in 3-, 10-, and 18-day old rat pups in isolation and the home cage. Developmental Psychobiology, 27: 65-83.