Erik Lee
@erikglee.bsky.social
Neuroimaging Researcher at the University of Minnesota and the Masonic Institute for the Developing Brain. Student in UMN's Bioinformatics and Computational Biology Program.
I like this one. They use invasive oxygen polarography to identify a band-limited peak (above background 1/f) around 0.06Hz — notably, this corresponds to the frequency range where they also observe the strongest functional connectivity. www.pnas.org/doi/abs/10.1...
Functional connectivity arises from a slow rhythmic mechanism | PNAS
The mechanism underlying temporal correlations among blood oxygen level-dependent
signals is unclear. We used oxygen polarography to better charact...
www.pnas.org
July 22, 2025 at 12:35 PM
I like this one. They use invasive oxygen polarography to identify a band-limited peak (above background 1/f) around 0.06Hz — notably, this corresponds to the frequency range where they also observe the strongest functional connectivity. www.pnas.org/doi/abs/10.1...
In the past I have run PCA on the input data first. Another option is to use a regularized variant of CCA that penalizes model complexity. I haven’t messed around with this much, but it looks like the CCA implementation in mvlearn supports regularization
October 25, 2023 at 8:25 PM
In the past I have run PCA on the input data first. Another option is to use a regularized variant of CCA that penalizes model complexity. I haven’t messed around with this much, but it looks like the CCA implementation in mvlearn supports regularization