Yuki Minai
@curi_ms
Followers
143
Following
344
Media
21
Statuses
459
Ph.D. student in Neural Computation and Machine Learning at Carnegie Mellon University Personal webpage: https://t.co/xe96gS5Me7
Joined August 2019
I'm excited to share our #NeurIPS2024 paper with.@jsoldadomagrane @SmithLabNeuro @YuLikeNeuro!.We develop a new brain stimulation framework (MiSO) to drive neural population activity toward specified states. Paper: Poster Session 4 East, Dec 12 16:30.[1/n]
4
39
181
RT @NeuroEngin_UMH: 🥳We are so glad to announce that our paper "Brain connectivity changes in response to cortical electrical stimulation i….
academic.oup.com
Abstract. The success of visual neuroprostheses in long-term blind individuals depends not only on the prosthetic technology but also on the brain’s abilit
0
3
0
Tomorrow, I will be presenting the closed-loop brain stimulation framework we developed (MicroStimulation Optimization, MiSO) at the workshop "Causal perturbation based approaches to uncovering neural dynamics". Looking forward to seeing you there! .#Coysne2025
0
4
27
RT @RaynorProject: We are excited to announce our 2025 RFA for $2.5 million Grant for an Artificial Intelligence Core for the newly launche….
0
5
0
RT @HunterSchone: After 10 years of implanting Utah Arrays in 5 participants with spinal cord injuries…. we provide a methodological roadma….
0
21
0
RT @giacomo_valle: Check out our new paper 'Evoking stable and precise tactile sensations via multi-electrode intracortical microstimulatio….
0
23
0
MiSO successfully searched amongst thousands of uStim parameter configurations to drive the population activity toward specified states. MiSO increases the clinical viability of neuromodulation technologies by enabling the use of larger stimulation parameter spaces. [7/n]
0
0
5
In this study, we implemented MiSO with a factor analysis based alignment method, a CNN model, and an epsilon greedy optimization algorithm. We tested MiSO in closed-loop experiments using electrical microstimulation (uStim) in the prefrontal cortex of a non-human primate. [6/n].
1
0
4
3) MiSO’s online optimization algorithm is then initialized with the model’s prediction, and subsequently updates the prediction and optimal stimulation parameters using the new observations in a closed-loop experiment. [5/n].
1
0
3
2) To learn the mapping between stimulation parameters and the brain's response given the small number of merged samples, MiSO’s statistical model leverages the structure in the brain responses to predict the responses of untested stimulation parameters. [4/n].
1
0
5
1) To create a stimulation-response sample dataset across sessions, MiSO’s alignment method merges neural activity across sessions in a way that is robust to recording instabilities such as baseline shifts. [3/n].
1
0
3
MiSO is a closed-loop stimulation framework to optimize over a large parameter space. MiSO consists of three key components: 1) a neural activity alignment method, 2) a statistical model capturing stimulation-response relationships, and 3) an online optimization algorithm. [2/n].
1
0
3
New blog series on Pokemon with RL :).
medium.com
In the previous blog, I introduced how to create a custom Gymnasium environment using Pokemon RED as an example.
1
0
0
RT @SmithLabNeuro: Come join us @cmuneurosci ! We're recruiting a new Assistant Professor in the Neuroscience Institute: .
0
18
0
RT @MaryamShanechi: New in @NatureNeuro, we present DPAD, a deep learning method for dynamical modeling of neural-behavioral data & dissoci….
0
83
0
"But do remember that your prime purpose is to explain something, not prove that you’re smarter than your readers.".
0
0
0