Ole Bialas

Research software consultant in the @IbehaveNetwork, trying to make neuroscience research more reproducible.

[Neuromatch @neuromatch]

💼 Paid Opportunity: Join Neuromatch Academy and Climatematch Academy as a #virtual #TeachingAssistant this July. 8hrs/day, Monday to Friday during the course dates.

➡️ Learn more here: https://t.co/osmQeVy1D4
➡️ Apply before 15 March https://t.co/t399SdP7NQ

[Ole Bialas @NeuroObi]

3. Reliance on US software is becoming a security risk. The increasing hostility drives European governments to seek open source alternatives.

[Ole Bialas @NeuroObi]

I think as long as it is still January, new year predictions are allowed, so here is one: 2026 may be the year of the Linux desktop 🐧. I can think of three good reasons: (read the full post here: https://t.co/c7lpIGEepT)

[Ole Bialas @NeuroObi]

1. Windows is on the decline. Bugs, bloat and unwanted AI features make it less and less usable
2. Linux has a lot of momentum. Adoption rates are rising and Linux increasingly establishes itself in the gaming market.

[Ole Bialas @NeuroObi]

On the day before, I'll host another 1-day workshop where you can learn the prerequisite computer skills like using a terminal and version control with git:
https://t.co/8ufq6ofdTM

[Ole Bialas @NeuroObi]

Want to learn how to manage and version-control scientific data using a free and open source distributed system?

Then join me on Dec 5th for a 1-day introductory workshop on DataLad!
It's online and free and you can sign up here if you are interested:
https://t.co/QjnewBFVWj

[Ole Bialas @NeuroObi]

I recently learned about UnitMatch, a Bayesian algorithm for tracking units across multiple recordings. There is a Python package but, unfortunately, I did not find any good tutorial on how to use it - so I made one! Hope it's useful.
https://t.co/3FzxBXpgq6

[Ole Bialas @NeuroObi]

New blog post! Using Python with Pandas and Seaborn to analyze a and visualize the activity from hundreds of neurons in the visual cortex:

https://t.co/IsFEfATyY0

[Ole Bialas @NeuroObi]

Want to learn new data analysis skills for neuroscience? Join my workshop "Intro to Neural Spike Analysis in Python" June 11th to 13th. It's free and online, and we'll explore interesting topics like model fitting, PCA and generative simulations
#Neuroscience #Python

[iBehave Network @IbehaveNetwork]

🔧New iBOTS Workshop! "Intro Neural Spike Analysis in Python"

📅June 11-13
Dr. Ole Bialas
open to all @UniBonn @UniklinikBonn @UKKoeln @UniCologne @MPINB_outreach @fz_juelich @RWTH
Learn how to process and analyze neural spiking data!
📌Register:  https://t.co/SIvTiqA4yw
#iBOTS

[Edmund Lalor @edmundlalor]

Are you Interested in:
- the role of oscillations in visual and auditory processing?
- blending computational and experimental neuroscience?
- research with clinical implications?

If yes, then maybe you'd like a postdoc with us.
DM/email me for more details.

RTs appreciated

[Ole Bialas @NeuroObi]

It's unnerving that the terms reproducibility and replicability are used interchangeably.
It shouldn't be hard:
If an animal REPRODUCES, it makes something that similar to itself but not identical.
If DNA is REPLICATED, the result is identical - if it isn't there was an error.

[Ole Bialas @NeuroObi]

Next week, I'll host a workshop on building robust neuroscience experiments in Python 🐍 and there are still some slots available. The course is Wed-Fr, 9:30am to 5pm CET It's completely free and online! Details in the registration:
https://t.co/XpxiLauqFp

[Ole Bialas @NeuroObi]

For a PsychoPy course, I wanted to show how loading images adds to the trial duration. On a new fancy notebook with Windows, loading took ~90ms while on an old notebook running Debian, loading just took ~30ms. Can't we just stop using Windows?

[Ole Bialas @NeuroObi]

Did you ever crash your experiment by messing up your configuration file? I know I have! In this blog post, I discuss how to use type hinting and data validation with @pydantic to avoid that this ever happens again:
https://t.co/SlNbmkDWxO

[Ole Bialas @NeuroObi]

Great opportunity to work in a really nice team - I miss that place a lot!

[Aaron Nidiffer @arnndffr]

Come work with me! (and other fine folks I guess)

RTs appreciated. Hit me up if you have any questions.

[Aaron Nidiffer @arnndffr]

Come work with me! (and other fine folks I guess)

RTs appreciated. Hit me up if you have any questions.

[Edmund Lalor @edmundlalor]

We are likely to advertise a postdoc position in the next month or so. If you are interested in brains, speech, language, oscillations, computational modeling, multisensory integration, predictive perception, please reach out and I can tell you more.

RTs appreciated.

[Ole Bialas @NeuroObi]

@matteo_brainnet @edmundlalor If the effects were caused by non-stationary trends in the data, I would expect that you get more variability as epoch duration increases

[Ole Bialas @NeuroObi]

New preprint with @edmundlalor where we show that non-stationarity impairs encoding models trained on continuous neural recordings. This can be remedied by appropriate segmentation of the data (for EEG, 10s seems like a reasonable default).
https://t.co/fq6k5cseAx

[Ole Bialas @NeuroObi]

@matteo_brainnet @edmundlalor Oh, that's interesting! I don't know much about connectivity measures, but I guess they assume sttaionarity as well. However, I'm not sure if that explains your findings. In your paper, the estimates appear to converge on some value as epoch duration increases.

[Ole Bialas @NeuroObi]

Code and data can be found here:
https://t.co/aaoS8cwmxU