iiinova — where your digital self keeps evolving

Daniel Borek

@danielborek.bsky.social

457 followers 1.5K following 76 posts

🇵🇱 PL in Brussels | A PhD candidate trying to make sense of human 🧠 #oscillations in #EEG #MEG using #R and #Python | other interests: knowledge management, #metascience, #OpenScience, #PhilosophyOfScience, #DataViz | 🎥 #cinephile

Posts Replies Media

Daniel Borek @danielborek.bsky.social · 25/05/2026

An XKCD-style webcomic featuring two characters against a yellow background. On the right, a wild-haired character in a lab coat stands next to a large mechanical device resembling a laser or death ray with a control panel. On the left, a blue-shirted character questions them.
The dialogue boxes read:
• Blue-shirted character: "Why did you build a death ray?"
• Lab-coat character: "To take over the world."
• Blue-shirted character: "No, I mean what mad hypothesis are you testing? Are you just making mad observations? You at least are going to leave some of the world as a mad control group, right?"
• Lab-coat character: "Look, I'm just trying to take over the world. That's all."
The caption at the bottom reads: "Sad truth: Most 'mad scientists' are actually just mad engineers".

010

Daniel Borek @danielborek.bsky.social · 23/05/2026

What happens when your specparam model specification doesn’t match your actual neural signal? I built a small WASM Python app to explore on simulations how model misfit can skew PSD shape and parameter estimates. Try it online here: danielborek.me/code-spectra... #EEG #MEG #Python #shiny

Screenshot of an interactive SpecParam knee simulation dashboard. The app shows controls for simulated spectral parameters, an explanation of knee parametrisation, and a log-log PSD plot comparing the true spectrum, noisy spectrum, knee model fit, and fixed 1/f fit. The knee model follows the curved spectrum, while the fixed model fit deviates as a straight line.

1116

Daniel Borek @danielborek.bsky.social · 21/11/2024

It is always interesting to look into the history of your discipline. After people started using Fourier transform for the spectral analysis of EEG (1932), in the late 30s and 40s many studies popped up looking for shifts in spectral peaks, almost like in the 2020s. Figure from Gibbs & Knott (1949)

130