Neuroscientist and musician. Currently getting my PhD at the lab of Prof. Loren Frank, University of California, San Francisco and UC Berkeley.
shijiegu.github.io
shijiegu
Loading...
2/ My thesis committee was skeptical about branching into this side project. But @akgillespie.bsky.social and I saw something cool here. (I had the luck to catch the last bit of Anna's time at the Frank Lab to iron out all the details about this sequence task.)
4/ In this way, we do not really need to start with SWRs or theta sequences, or any prior about neural data. We will start with behavior, and we can be open to all the possibilities in neural data.
3/ One of the hardest problems sometimes is knowing *when* to look for thinking. I can record 8 hours of neural data a day — but where in those 8 hours should I analyze? COM gives you a natural behavioral marker: a clear time window when deliberation is likely happening.
7/ During my PhD, I have worked on many projects. There is one on real-time neural data-triggered optogenetics, which is now passed down to other lab members. We can quench neuronal firings 10s of milliseconds after the onset of a SWR event (bouts of strong neuronal activity in the hippocampus).
5/ What we found with hippocampal ephys recording: COM reflects a 2-stage process — first, detecting a likely error; then, generating and evaluating alternatives before committing to a revised choice.
1/ I stumbled onto this accidentally while studying rats' foraging biases. When I set up a task that forced rats to override their instincts, they'd spontaneously show COM: walking down a corridor, stopping, backing up, then choosing a different path.
[I was also working on a 3-stage RISC-V CPU design for my EECS251A course FPGA project at the time, so maybe this "2-stage process" of packaging my paper is a little inspired by that?]
6/ The hippocampus appears central to the second stage and not the first. We can predict the animal's next decision from hippocampal activity well before the choice is made.
Last week, we finally submitted my first paper during my PhD (bioRXiv: www.biorxiv.org/content/10.6...). The paper is about change-of-mind (COM) behavior.
You know when you're taking a test and switch your answer after second-guessing yourself? We studied what happens in the brain at that moment.
8/ Very fortunately, this (also side-) project got me in conversation with Blackrock Neurotech, and I am doing an internship there this summer. I am so excited to be in SLC and be back in software development for real-time systems.
