iiinova — where your digital self keeps evolving

David Haydock @davidghaydock.bsky.social · 19/11/2025

We collected ~4.7 hours of task data on 3 participants, listening to 10 different categories to create a functional gradient space in which we could situate category structure, and infer acoustic and semantic dimensions. Oh also, the data is open access! 👉 openneuro.org/datasets/ds0...

openneuro.org

OpenNeuro

031

David Haydock @davidghaydock.bsky.social · 19/11/2025

🧠🔊 New preprint! How does the brain organise natural sound categories? In our new fMRI study, we show that auditory cortex arranges sounds along continuous functional gradients, where acoustic and semantic structure are integrated into a shared representational space. doi.org/10.1101/2025...

doi.org

Auditory Cortical Gradients Integrate Bottom-Up and Top-Down Structure During Natural Sound Categorisation

Understanding how the brain organises natural categories is a central challenge in neuroscience. While prior work has shown that categories can be decoded from distributed activity patterns in auditor...

132

David Haydock @davidghaydock.bsky.social · 20/03/2025

We really need to start naming things in neuro stuff like "the sonic hedgehog pathway"

000

Reposted by David Haydock

Jonny Smallwood @themindwanders.bsky.social · 24/02/2025

This looks great! Excellent work by @davidghaydock.bsky.social et al.

013

David Haydock @davidghaydock.bsky.social · 21/02/2025

Our review provides a roadmap for researchers working on EEG microstate syntax. We hope to make results more comparable and useful for future studies, and call on researchers in the field to associate microstates to a continuous signal.

100

David Haydock @davidghaydock.bsky.social · 21/02/2025

Beyond this, we point out that existing methods which try to associate EEG microstates with fMRI patterns make sweeping assumptions when using GLM models by averaging the EEG time series to single values per TR, heavily simplifying the EEG signal.

100

David Haydock @davidghaydock.bsky.social · 21/02/2025

We argue that this common criticism can be investigated without throwing away microstate analysis by studying microstates in a continuous space (such as a t-SNE embedding, or similar). Note the information lost in the microstate representation!

101

David Haydock @davidghaydock.bsky.social · 21/02/2025

A key issue highlighted (among others): is how microstate sequences are generated in and of themselves, in a “winner-takes-all” approach, where the complexity of the continuous EEG signal is simplified to a sequence of symbols.

100

David Haydock @davidghaydock.bsky.social · 21/02/2025

In our new review, we organise existing methods into clear categories and define how different studies construct microstates, define microstate sequences, and how they go about investigating a sequence once they have it.

100

David Haydock @davidghaydock.bsky.social · 21/02/2025

Studies on microstate syntax use a lot of different methods, and don’t always use the same process for defining the microstate sequence. Different terms are used for the same concepts, and documentation of the specifics of preprocessing and analysis steps can be lacking.

100

David Haydock @davidghaydock.bsky.social · 21/02/2025

Ever heard of EEG microstates? Usually defined as cluster centres of topography, the dynamics of MS sequences are referred to as "syntaxes". Our new review discusses syntax methods and show how they could be better associated with the underlying EEG signal: 🧵👇 www.sciencedirect.com/science/arti...

164

David Haydock @davidghaydock.bsky.social · 26/12/2024

Further than that actually - mind and world, perception and world, are just as inseparable.

000

David Haydock @davidghaydock.bsky.social · 26/12/2024

Our perception and the world aren't two separate things where one is guessing about the other - they're unified aspects of a single lived experience that co-emerge and define each other.

000

David Haydock @davidghaydock.bsky.social · 26/12/2024

No disrespect but saying that perception of the world is a brain-based best guess is like saying that a wave is the ocean's best guess about what water should do. Brain and world aren't separate things making predictions about each other - they're inseparable aspects of the same dynamic process.

200

David Haydock @davidghaydock.bsky.social · 15/12/2024

If you'd have written an original piece on the matter and given inputs and opinions then there would have been a healthy discussion of the subject. Instead you've just created an argument about fair use and consent. Both of which you seem to be on the wrong side of.

000

David Haydock @davidghaydock.bsky.social · 15/12/2024

This isn't audience reaction, it's *author* reaction. That discrepancy is why you can't act as if use of LLMs is the same as writing an original piece. In any other context where you wrote about the paper, the contents of the paper is what would be discussed, which is what journalism should be for.

200

David Haydock @davidghaydock.bsky.social · 15/12/2024

It's one thing to summarise it for yourself with a language model to get the gist of an article, but it's another thing entirely to re-present the article publicly using that summary.

110

David Haydock @davidghaydock.bsky.social · 15/12/2024

It makes more sense to me to identify the parts you'd want to run yourself. Relying on the summary of a language model instead of taking the time to absorb the source material inevitably means you will miss the finer detail. Like reading the abstract of a paper and thinking you understand the whole.

120

Reposted by David Haydock

🌙 Lune Bellec @lune-bellec.bsky.social · 06/12/2024

Picking a journal to publish our work sometimes felt like a headache. High fees and private, for-profit governance remain the (terrifying) norm. But this is changing: Aperture is a recent neuroimaging journal that is community-driven, low cost, and strives for high quality. See ⬇️

074

David Haydock @davidghaydock.bsky.social · 05/12/2024

Open science includes cool animations then

100

David Haydock @davidghaydock.bsky.social · 05/12/2024

Also, this needs some Dorian Concept as background music

020

David Haydock @davidghaydock.bsky.social · 05/12/2024

You need to put a watermark on these before someone nicks them and starts using them in TikTok's that have nothing to do with brain science

220

Reposted by David Haydock

Micah G. Allen @micahgallen.com · 17/11/2024

IMO the big breakthroughs won’t just come from advancing temporal and spatial resolution, but from achieving naturalistic recordings during dynamic, embodied interactions. Until we can move away from highly restrictive and overly reductive passive stimulation we won’t fully grasp mechanism.

8603

Reposted by David Haydock

Robert leech @leechbrain.bsky.social · 24/09/2024

Brain video 25b: #blender #blender Fight the scanner: part 2

021