Above: A delightfully old picture of the author giving a poster presentation in Budapest at the Human Brain Mapping conference, Summer 2004. As should be apparent from the gleam in the author's eye, this is before he spent 6 hours explaining the material again and again to attendees with widely varying degrees of interest.
The Vanderbilt Computational Memory Lab, directed by Sean Polyn, is
part of the Vanderbilt University Department of Psychology.
Our lab is interested in the cognitive and neural dynamics of the human memory system, and more specifically, how we use this system to search through our memories of recently learned material. Every day, we store hundreds of new memories; sometimes these memories can be retrieved and examined effortlessly, but sometimes, to our frustration, we find our efforts blocked, and our memories inaccessible. The brain contains sophisticated neural machinery allowing us to target particular memories. How does this machinery work, and why does it fail? We believe in a multi-tiered approach to the study of human memory, combining neurorecording techniques (fMRI and EEG), with behavioral investigations and computational modeling. These multiple levels of analysis inform one another, and allow us to constrain our understanding of human memory.
Email me at: firstname.lastname@example.org
Sean Polyn's Curriculum Vitae: (PDF)
For a full list of publications, please refer to the Vanderbilt Computational Memory Lab's publications page.
Polyn S. M., Norman K. A., and Kahana M. J. A context maintenance and retrieval model of organizational processes in free recall. Psychological Review, 116(1), 129-156. (PDF)
Kahana, M. J., Howard, M. W. and Polyn, S. M. Associative retrieval processes in episodic memory. In H. L. Roediger, III, editor, Cognitive psychology of memory. Vol. 2 of Learning and memory: A comprehensive reference, 4 vols. (J. Byrne, Editor). Elsevier, Oxford, 2008. (PDF)
Polyn S. M. & Kahana M. J. (2008) Memory search and the neural representation of context. Trends in Cognitive Science, 12(1), 24-30. (PDF)
Norman K.A., Polyn S.M., Detre G.J., & Haxby J.V. (2006) Beyond mind reading: multi-voxel pattern analysis of fMRI data. Trends in Cognitive Science, 10(3), 424-430. (PDF)
Polyn S.M., Natu V.S., Cohen J.D., & Norman K.A. (2005) Category-specific cortical activity precedes recall during memory search. Science, 310, 1963-1966. (link to paper)
Polyn S.M. (2005) Neuroimaging, behavioral, and computational investigations of memory targeting. Unpublished doctoral dissertation. Department of Psychology, Princeton University. (PDF)
Polyn S.M. (2003) Connectionist modeling of context change effects in recognition memory. Unpublished Master's thesis. Department of Psychology, Princeton University. (DOC)
Polyn S., Wu X.B., & Levy W.B. (2000) Entorhinal/dentate excitation of CA3: A critical variable in hippocampal models. Neurocomputing, 32-33, 493-499. (PDF)
Polyn S.M., Norman K.A., & Kahana M.J. (2008) Context maintenance and retrieval: A model of episodic and semantic organization in free recall. Spoken presentation at Society for Mathematical Psychology meeting. Washington, D.C. (ZIPPED KEYNOTE)
Miller J.F., Polyn S.M., & Kahana M.J. (2007) Clustering by spatial proximity during memory search. Poster presented at Society for Mathematical Psychology conference. Irvine, CA. (PDF)
Polyn S.M., Nystrom L.E., Norman K.A., Haxby J.V., Gobbini M.I., & Cohen J.D. (2004) Using neural network algorithms to investigate distributed patterns of brain activity in fMRI. Poster presented at OHBM conference. Budapest, Hungary. (PDF)
Norming materials - Task word pool (TGZ).
The task word pool contains 1297 words; each of these words has been judged by 12 subjects using three tasks: Pleasantness, Size, and Animacy. The statistics regarding these judgments are included, along with a technical report describing the details of the norming experiment.
Detection of retrieved memories: A major component of my PhD work was a neuroimaging study of human memory. A quick summary: We measured people's brain activity while they studied a number of things, as well as when they later tried to remember those things. We were able to tell what type of thing they were trying to remember, often several seconds before the person started reporting specific items! This work was published in Science Magazine in the 23 December 2005 issue ("Category-specific cortical activity precedes retrieval during memory search"). See above for a link to the article.
Some press from the memory-retrieval paper:
An interview I did with the BBC program "Science In Action". (mp3)
An article written by Bloomberg News. (txt)
An article that appeared in Seed Magazine. (link)
An article that was up on MSNBC. (plain txt)
An article that is up on WebMD. (link)
Connectionist model of human memory: As one part of my PhD work, I developed a connectionist model of prefrontal and medial temporal interactions; the model is an attempt to describe the memory search process that occurs during episodic memory tasks such as free recall. The model is a modification of one described by Ken Norman & Randy O'Reilly in an article in Psychological Review (2003). My modifications are inspired by Mark Howard & Mike Kahana's model of Temporal Context Memory.
The gist of the theory: Humans maintain a pattern of activity in their brain that generally reflects the details of what has been going on recently (also known as "spatiotemporal context" or just temporal context). Whenever a new memory is stored, it includes within it this temporal context pattern. That means when the memory is retrieved, the temporal context pattern is retrieved as well. Coming soon: A pictorial tour of the model. For more detail, consult my dissertation, linked above.
Classification of distributed patterns in neuroimaging data: The rest of my time is spent caressing fMRI data acquired with the 3T Siemens magnet located in Green Hall at Princeton University (and maintained by the Center for the Study of Brain, Mind, and Behavior). The memory study described above used these techniques.
Detection of categorical representations: This work was most recently presented at the Human Brain Mapping conference in Budapest, Summer 2004. In collaboration with Jim Haxby, Ida Gobbini, Ken Norman, Jon Cohen, Youssef Ezzyat & Leigh Nystrom. Subjects in the scanner view pictures of stimuli drawn from a number of categories: male faces, female faces, monkey faces, dog faces, houses, shoes and chairs. Using neural network classifiers (backpropagation), we are able to identify the category of stimulus being viewed with near ceiling accuracy.
Computational Memory Lab
Ken Norman: Princeton Computational Memory Lab
Jon Cohen: Neuroscience of Cognitive Control Laboratory
William Levy: Laboratory of Systems Neurodynamics
Christoph Weidemann: Scientific Wunderkind
Shallow Grave Comics
Eric Polyn: San Diego Underground Music