Sharp-Wave Ripple (SWR) complexes are increases in neuronal oscillations in the ripple band (150-250 Hz), typically correlated with sharp-waves in the stratum radiatum. These SWR complexes are crucial for memory consolidation — they are strongly correlated with the sequential reactivation of specific cells within the hippocampus, typically in the same order they fired during learning. As one may guess, disruption of such events during learning or sleep may prevent the animal from consolidating the events that transpired during the training sessions, effectively blocking the memory. However, SWR complexes are fairly quick events (~100 ms on average). Though they can be detected by current methods with a fairly high true positive rate, the issue comes into play when one attempts to attain a degree of specificity in stimulation. Depending on how the event boundaries are defined, 40-60% of the ripple event is likely to be missed by stimulation.
Shark-Wave attempts explores machine learning methods for time series predictions, with the goal of uncovering predictive features for SWR complexes in rodents. I am leading this project, and working in collaboration with Shayok Dutta and James Webb on this project. Due to the use of private data from another lab, I cannot use a public GitHub repository for this one, and thus you are link-less in terms of access to my code. However, I’d be happy to answer any and all questions on the project that I can disclose!