Lightnews — Scholar-powered news

Reposted by Athena Akrami

Carlos Brody @carlosbrody.bsky.social · 19d

What happens in your brain when you make up your mind?

Postdoc (soon faculty at U. of Utah) @thomas-zhihao-luo.bsky.social and ex-grad student (now Shanahan Fellow at Allen Institute) @timkimd.bsky.social have some answers in this new paper out in Nature!

www.nature.com/articles/s41...

🧵 1/6

1 11 50

Athena Akrami @athenaakrami.bsky.social · Jul 20

Thanks, Matteo! Will do.

Reposted by Athena Akrami

Matteo Carandini @carandinilab.net · Jul 20

Folks in the UK who want to help may want to write to their MP. The list is at members.parliament.uk/constituencies. (I wrote to mine).

Constituencies - MPs and Lords - UK Parliament

Search and find constituencies in the United Kingdom by name, postcode or location.

members.parliament.uk

1 1 1

Athena Akrami @athenaakrami.bsky.social · Jul 20

Without paywall link:
archive.is/678If

1 1 6

Reposted by Athena Akrami

Kevin Mitchell @wiringthebrain.bsky.social · Jul 19

How long is the world going to allow this horror to go on?

Sky News @news.sky.com · Jul 19

Gaza: More than 30 people killed 'as Israeli troops open fire towards Palestinians waiting for aid'

More than 30 people have been killed after Israeli troops opened fire towards crowds of Palestinians waiting for aid, according to witnesses and hospital officials.

news.sky.com

6 35 99

Athena Akrami @athenaakrami.bsky.social · Jul 18

And read the excellent post by the amazing science communication team at @sainsburywellcome.bsky.social, explaining "Heron the progeny of Bonsai and ROS".

15/
sainsburywellcome.org/web/blog/dev...

Developing Heron, the progeny of Bonsai and ROS | Sainsbury Wellcome Centre

sainsburywellcome.org

2

Athena Akrami @athenaakrami.bsky.social · Jul 18

Read George's blog post on why/how he developed Heron:
python.plainenglish.io/heron-a-hybr...

14/

Heron: A Hybrid Approach to Data Pipelines in Python

Concept and a pre-manual

python.plainenglish.io

1 1

Athena Akrami @athenaakrami.bsky.social · Jul 18

The code for all repositories can be found on GitHub:
github.com/Heron-Reposi...

Access Heron’s documentation here:
heron-42ad.readthedocs.io/en/latest/

13/

Heron-Repositories

Heron-Repositories has 14 repositories available. Follow their code on GitHub.

github.com

1 2

Athena Akrami @athenaakrami.bsky.social · Jul 18

Heron is open source, and we'd love to get feedback from the wider Python community.

We'd also like to explore how Heron can be used on other data analysis pipelines (besides neuroscience), and hope to see a community forming around the use and update of the code.

12/

1 1

Athena Akrami @athenaakrami.bsky.social · Jul 18

Heron is for those who
- Demand full control of their setup

- Prefer not to have to choose bet hardware/software that run only on a specific chip/OS combo

- Appreciate the ease & speed that high-level languages (eg Python) & Graphical User Interfaces (GUIs) offer them.

11/

1 1

Athena Akrami @athenaakrami.bsky.social · Jul 18

Heron is designed from the ground up to allow every experiment to be implemented using the tools the developers are more familiar with & the hardware they happen to own in the lab.

10/

1 1

Athena Akrami @athenaakrami.bsky.social · Jul 18

Heron’s main principle is to allow researchers to design and implement the experimental flow as close as possible to their mental schemata of the experiment, in the form of a Knowledge Graph. Self documentation and ease of reconstruction/reconfiguration comes for free.

9/

Heron’s combination of graphical and text-based code development.
(A) Schematic of Heron’s two levels of operation. Left shows a Knowledge Graph made of Nodes with Names, named Parameters and named Input/Output points. Nodes are connected with links (Edges) that define the flow of information. Right shows a piece of textual code that fully defines both the Node’s Graphical User Interface (GUI) appearance and the processes run by the machine. (B) Heron’s implementation of A, where the Knowledge Graph is generated in Heron’s Node Editor window while the code that each Node needs to be defined and the code that each Node is running is written in a separate Integrated Development Environment (IDE, in this case PyCharm).

1 1 4

Athena Akrami @athenaakrami.bsky.social · Jul 18

These all motivated George to develope Heron, a Knowledge Graph editor, that allows for hybrid programming. i.e. it has a graphical interface, which behaves a lot like Bonsai, but users can write their own nodes in Python.

8/

1 1

Athena Akrami @athenaakrami.bsky.social · Jul 18

Another issue to consider is to fit one's mental image of the experimental pipeline into existing commercialised or customised platforms.

Don't you like to be able to design & code up your experiment, as close as possible to your "mental schema" of how it would unfold?

7/

1 1

Athena Akrami @athenaakrami.bsky.social · Jul 18

Bonsai, is a tool that is very easy to use and put together simple experiments. However, it gets more difficult as the complexity of experiments increases.

6/

A graph showing how 'difficulty of use' scales with 'experiment complexity'.

1 1

Athena Akrami @athenaakrami.bsky.social · Jul 18

You may think, but wait, don't we have already a few pretty good options, like Bonsai & ROS?

Yes, they're powerful and great tools, but the ROS is mainly built in C++ and so it is not an easy tool for many experimentalists to use.

5/

1 1

Athena Akrami @athenaakrami.bsky.social · Jul 18

Some platforms make it very difficult to flexibly connect various hardware that may run on different operating system/chip combinations. Or, piecing codes from different programming languages.

4/

1 1

Athena Akrami @athenaakrami.bsky.social · Jul 18

Some are too low level, too complicated. Some are too black-boxy, too opaque.

Some will end up with a spaghetti code, you have no idea how to keep track of after few iterations, let alone other ppl who'd like to use & expand your code/design.

3/

1 1

Athena Akrami @athenaakrami.bsky.social · Jul 18

Those of you who've attempted to code up a new experiment, requiring various complex hardware and software, know very well how daunting it can be to pick the right system-design platform and development environment.

Some are too specific, too restricted.
2/

1 2

Athena Akrami @athenaakrami.bsky.social · Jul 18

🎉 Heron is finally out @elife.bsky.social! Led by George Dimitriadis, with Ella Svahn & @macaskillaf.bsky.social

🧪 🧠 🐭 🤖

If you wonder why yet another tool for experimental pipelines, read the 🧵 below:

#neuroscience #neuroskyence #OpenSource

1/
elifesciences.org/articles/91915

1 12 38

Reposted by Athena Akrami

eLife @elife.bsky.social · Jul 14

Archerfish make lightning-fast decisions to catch prey, but their aim isn’t hardwired. They can adapt to new physics and generalise rules across contexts, all in under 100 ms.
buff.ly/7Ogh3B1

1 9

Athena Akrami @athenaakrami.bsky.social · Jul 8

I fully agree -- RNNs do form attractor dynamics. That cannot be used as an argument against attractors ...

1 3

Athena Akrami @athenaakrami.bsky.social · Jul 8

2\ In rodents, many important works from Svoboda (in collaboration w Romani) lab on motor preparation in ALM/thalamus get into nitty gritty of proving attractor dynamics, with most of your check list ticked

2 3

Athena Akrami @athenaakrami.bsky.social · Jul 8

Besides the misleading title, I disagree with Konrad on the RNN case -- training an ugly RNN to show all the properties of attractor dynamics means that the RNN has formed attractors!

1 1 10

Athena Akrami @athenaakrami.bsky.social · Jul 8

1\ Reading the title (as well as opening w 'mathematicsl objects can not be') sounded as if you're suggesting attractors, conceptually & in essence, cannot be considered as mechanistic models. But what you're saying is that experimentally, things have been consistent w attractors, but not full proof

1 2