We hosted the 27th Annual Meeting of the Center for Neurobiology of Learning and Memory at UC Irvine last week and the topic was (drumroll please…..) you guessed it. Pattern separation! While devoting a whole 1.5 day conference to such a “narrow topic” may not have seemed like the best idea, somehow it worked. In fact, most attendees commented that the focused format allowed for very interesting discussions. As always, there’s never enough time for discussion, but that’s what this forum is for.
In the first session, we heard from Bruce McNaughton on modeling the allocation of memories to different neurons in the hippocampus, from Jill Leutgeb about orthogonalization in the dentate and potential mechanisms supporting pattern separation and from Jim Knierim on pattern separation and completion in the EC-DG-CA3 circuit. The discussion here, led by Ivan Soltesz, focused on granule cell properties, the architecture of the dentate gyrus, and what kind of evidence is necessary to conclude that pattern separation occurred.
In the second session, we attempted to make the links to behavioral outputs. We heard from Mike Fanselow who discussed the role of adult, newborn, and immature granule cells in contextual fear discrimination as well as generalization, from John Guzowski on measuring the impact of neuroinflammatory response on context discrimination memory and pattern separation, and finally from Craig Stark on inferring pattern separation in the human hippocampus using a combination of high-resolution fMRI and cognitive tasks. I led the discussion here and focused mainly on the relationship between behavioral tasks and neural pattern separation. Opinions were bipolar on this issue, with some folks agreeing that links need to be made and others saying that behavior is far removed from pattern separation. This is the type of discussion we need to continue on this forum.
In the third session, we learned about the role of neurogenesis in the dentate. We heard first from Brad Aimone who dissected the evidence for and against his model of newborn granule cell involvement in pattern separation. A revised, biologically plausible model (thank you Sandia supercomputers) was better able to explain why loss of newborn cells might impair pattern separation despite their heightened excitability. Paul Frankland then discussed the role of adult neurogenesis in pattern separation using an IEG approach to look at population dynamics and presented some fascinating new data on forgetting (which I won’t share here). The discussion was led by John Guzowski and focused on properties of newborn granule cells and their potential roles in pattern separation. Some questions as to whether recordings in the dentate wouldn’t preferentially pick out these cells due to their hyperexcitability.
In the fourth and final session, we heard from Norbert Fortin who (despite being a self proclaimed “non-pattern-separation-guy) educated the crowd on similar architectures to the hippocampus and the dentate in particular in other species such as birds and reptiles. The evolutionary approach was very much appreciated and provided for interesting discussion. The last talk was delivered by Don Wilson who discussed the evidence for pattern separation in the olfactory system and opened up more avenues for discussions related to this phenomenon being as global as action potentials.
During our final discussion, we debated several issues including the merits and pitfalls of using discrimination tasks, whether we should be using behavior to infer something about computation and the caveats involved, whether pattern separation is a helpful concept if it’s so general, how to define it neurally or computationally, and whether we have yet to observe direct evidence for the computation (we certainly see evidence for its output, but the actual computation is likely synaptic and has not yet been observed).
The conference ended with more questions left than answers… but that’s the fun of it.
That’s my summary. You can see the entire program of course here. Feel free to ask questions and stir up discussion.
Tim Bussey and Ryan Hunsaker, we missed you guys at this meeting. Would love to hear from more behavioral neuroscientists on some of these issues. Get ready for more soon!