Redish Lab

Scientist studying learning, memory, and decision-making. Poet and Playwright.

Lab Website: https://redishlab.umn.edu/
Books: Changing How We Choose: The New Science of Morality
Books: The Mind within the Brain

Public Key

npub16jr6tn76pth8u36wjh0j09zmed42aue7slg8er9nw3qtl6wuh3us8cknep

NIP-05 Address

adredish_at_neuromatch.social@momostr.pink

Profile Code

nprofile1qqsdfpa9eldq4mn7ga8fthe8j3duk64w7vlg05ru3jehgs9la8wtc7gprpmhxue69uhhyetvv9ujumt0d4hhxarj9ecxjmnthgjvh0

Publishing to

relay.momostr.pink

Author Public Key

npub16jr6tn76pth8u36wjh0j09zmed42aue7slg8er9nw3qtl6wuh3us8cknep

Show more details

Last Notes

2026-07-18 10:52:09 UTC

- reply

💯 Even if the GPT model provides a valid reference, if you don't engage with it, you are not doing scholarship.
When I teach students, I tell them that they should never cite anything they haven't actually read, or even better read and re-read.
It's very disappointing (although not surprising) to find that the editors of these GlamourMagz are doing bad scholarship. (Most of the click-bait news-journalism parts of these magazines has been crap for decades and has actively damaged the scientific enterprise.)

2026-05-16 11:54:31 UTC

An absolutely fascinating article on the self-destruction of the American Hegemony and the implications for software technology around the world.
Found from @npub1fdr…lvhs 's recent blog post. (Which is always worth reading.)
https://www.baldurbjarnason.com/2026/the-old-world-of-tech-is-dying/

2026-03-05 13:08:01 UTC

- reply

One of my colleagues (Steve Kelley) had a very helpful perspective on this. He said that one should always engage with politicians (and other activists) knowing which of three hats you were wearing:
Hat 1. This is my expertise. I am the world expert in this topic. It comes from my research and I am *the* person to talk to about this topic.
Hat 2. This is my field. I know the nuances and questions that one needs to ask. I'm qualified to tell you which science is well-founded and which is bogus.
Hat 3. I'm a citizen. I'm a reasonably smart person who wants to share my opinions on what one should do in this situation.
I've found that talking to politicians being explicit about which hat I'm wearing has been very helpful. (And all three hats can be important. Remember, politicians want to know what their constituents want ... well, good politicians do. This why "call your representative" matters.)
Steve knows from what he is talking about. He was a state senator here in MN and then transitioned to be faculty in our School of Public Affairs. He ran a class once for other faculty in engagement, which is where I got this from.

2026-03-02 15:56:28 UTC

- reply

I think this is a good idea.
Anyone who wants a grant has to do at least 3 reviews for each submission. Anyone who wants to publish a paper has to do at least 2 reviews for each submission.
Not only would that help the review backlog, it would help limit the people who submit way too many grants and papers.

2026-01-29 13:16:37 UTC

- reply

That perspective makes me terribly sad. Science should be about communication and community. Tell everyone about all of your ideas because their dialogue will make your ideas better.
Moreover, there is room for all of us. We need to remember that the problem isn't that there isn't enough science funding for all the positions. The problem is that we have given all of our money (including all of our science money) to a handful of billionaires. (Apparently so that they can fuck up our politics and our society.)
And before someone says "there's no political will for this", I would point out that political will changes over time. Don't give in to their arguments.
If you're going to fight for something, fight for enough science budget that anyone who wants can do the work that helps the world be better. Don't fight to kick people out of this community just because they are older (or because of their skin color or their gender or any other unrelated-to-the-science-they-are-doing characteristic).

2025-12-30 23:22:14 UTC

Found out my kid uses #spotify I want to recommend an alternative. What's the best alternatives to suggest to them? thx.
Adding #BoycottSpotify because I assume people following that hashtag will know other options.

2025-12-06 13:13:59 UTC

- reply

This is classic corporate responses to academia.
I once was visiting London for a conference, mentioned this to a former student who was (at the time) working at DeepMind (I think it was part of Google then, but don't remember). He suggested I give a talk, and I said sure. They wanted me to I sign an NDA as "standard practice when someone visits DeepMind".
I told them I do not sign NDA's on principle. (I don't believe they are ethical. Also, I'm terrible at keeping secrets. :) I told them that I was on the outside and that anything they told me was public. They said "then you can't meet our people". I told them that was their loss not mine.
PS. The eventual outcome was that they did have me give the talk. I met lots of great people and we had great conversations about both the work I do and what they were working on. So ¯\_(ツ)_/¯. No, I did not sign an NDA.

2025-12-03 15:17:53 UTC

ICE is apparently in Minneapolis raiding buildings.

2025-09-07 21:19:04 UTC

[@knutson_brain](https://sfba.social/@knutson_brain )
You were right and I was wrong.
#Wikipedia works because it has a central set of sources that it accepts. One can only add a fact if there is a source that is real, cite-able, and reliable. Moreover, what Wikipedia promises is not "X is true", but rather "Z says X is true". This must be verifiable and Z must be a reliable source.
So Wikipedia does have a restriction process like peer-review (although it is different).
See the excellent (as always) discussion by @npub1fdr…lvhs: https://pluralistic.net/2025/09/05/be-the-first-person/
1/2

2025-08-13 13:09:08 UTC

There were two things that made America the economic powerhouse that it is. One was the innovation and hard work we got from immigration. The other was the enormous federal spending on science independent of its potential immediate relevance, which generated novel innovations. The key to both of these is the huge long-term gains, which manifest over the time-line of decades.
The current administration couldn't be doing more damage to the US if they actually were traitors and quislings.
😢

2025-08-02 18:22:01 UTC

An interesting thing that I realized in a response to @npub1xne…xqch and @npub14c5…rfvf.
There are ~60k postdocs in the US. Assume a lab costs on average $500k/yr. 60k PD @ $500k/yr = $30b.
Giving EVERY current postdoc in the US a lab would require doubling the NIH budget, but not more than that. 🤯
We just gave $300b/yr in tax cuts to the wealthy. 🫤
I don't know EU numbers, but I assume they're similar.

2025-07-17 13:26:11 UTC

- reply

I think it's also important to point out that these three techniques (agree - all are important!) sample different aspects of neural signalling. It's not just fidelity or scale.
Patch clamp measures in vivo voltage fluctuations, but those are not necessarily transmitted down axons, and you can have axonal generation of signaling. Electrode arrays (wires, tetrodes, silicon arrays, neuropixels) measure spikes (and LFPs, which are also a different signal, providing different information about the system). Calcium imaging provides information about calcium levels.
Although these signals are correlated with each other, they are not identical and should *never* be substituted one for the other. You can have calcium without spikes, synaptic interactions (intracellular signals) that do calculations, spikes without calcium, etc etc etc.
Instead of asking "can we see replay with calcium signaling?" I think we should ask "what is the calcium signaling doing when hippocampal sequences are replaying?"
That's why we need a literature. That's why we need all of these techniques. The information is synergistic.

2025-07-11 20:14:26 UTC

It's amazing how different economics looks when you assume the goal is to have an EFFECTIVE system rather than an EFFICIENT system.
An efficient system uses resources maximally, thus resources are never left lying around, but needs must be waiting for resources to become available.
An effective system resolves needs maximally, doing its best to ensure no needs are left unfulfilled. This requires that resources must be kept idle for unexpected needs.
In practice, effective systems are often more long-term efficient because emergency needs are often particularly expensive.

2025-03-14 15:52:18 UTC

- reply

Why were non-US arXiv mirrors shut down?
what about the other rXiv's (bioRxiv, medRxiv, psyRxiv)?

2024-12-04 13:39:19 UTC

As usual, @npub1w88…f8j6 misses a key factor in the story. While I agree with all of the points raised in this story (https://www.nature.com/articles/d41586-024-03906-0), it assumes that the government entities have the same GOAL as the science provides. But we know that many people in power do not have this same goal. We know that one of the reasons scientists are not asked to provide advice to governmental entities are that they give answers that people in power don't want to hear.
The survey in Figure 2 includes "politicians don't understand science" and "misinformation obscures science advice", but it doesn't include "politicians don't want to achieve the goals scientists provide".
Remember, the energy companies correctly predicted climate change (accurately! https://www.science.org/doi/10.1126/science.abk0063). It's not that they didn't understand the science. It's that they didn't want the outcome it would drive them to.
Note: unfortunately, half of the article is paywalled, but I got through to the rest of it from my institution and there's nothing particularly new in the last half.

2024-06-30 16:39:28 UTC

- reply

Re: interested vs disinterested evaluation.
When I want to know whether to believe some result and whether to act on it or not, I really want that evaluation to be from a party only interested in getting the right answer and not from someone with a conflict of interest (either pro or con).
If I want to know whether to plan for rising seas when I build my house, I want a geo-climate expert, not a debate between a climate expert and a climate change denier. If I want to know whether to vaccinate my kids, I want to get my information from NIAID, not from RFK Jr., even though he is apparently a very interested party. The correct person to assess whether a plane is safe to fly is not the manufacturer nor its competitors. It's the FAA.
Similarly, when I want peer review to assess my work (or someone else's), I want a disinterested party. Sure, as a scientist, I like to be challenged, and certainly I've had my share of fun in response to "adversarial collaborations", but it is definitely better when these are more "collaboration" and less "adversarial".
I do think it is sometimes fun to see two scientists really going at it in highly interested open discussion. But I think a lot of that is about seeing a good sporting match. In my experience, this kind of mano-a-mano fight is best done in published papers back and forth, with _external_ (disinterested!) reviewers assessing each paper (like a referee).
I don't think a mano-a-mano fight is evaluation. I think it is scientific discussion, which is different.
In my experience, the best *reviews* have always definitely more been experts interested in checking my answers, not someone particularly interested in proving me wrong.

2024-06-29 21:26:32 UTC

- reply

The true thing is peer review, that is, non-conflicted experts assessing the validity of the work.
Peer review can be improved. We (you, me, and many others) have spent years discussing various ways to implement it.
The journal was invented in 1665, where scientific editors judged quality ("peer" review of the work) or societies judged entry (review of the individual, but not the science itself). Assessment cycles in which experts provide advice to those editors and authors respond was invented in the 1950s. Frontiers developed an endorsement model that reduced gate-keeping and editor power in 2007. eLife started using its preprint commentary model in 2022 that enhanced editorial power, but was post-preprint peer review. We're still working on the best solution.
But... the most important thing is peer review is not some random comments from someone (often in conflict of interest) who has something to say about the paper. It is an assessment by an expert that takes time and effort and is done for others, not for oneself. It is a very powerful tool to improve scientific progress.
At this point, my opinion is the following (but I am still observing the scientific sociology to see how things work):
- Within field, preprints work fine.
- The binary published/not of "peer reviewed" journals did not work as well as we hoped.
- For scientists outside of a field, the new #eLife system of commentary works well. It provides detailed commentary that one can use to judge a paper's quality.
- For non-scientists (journalists, policy-makers, judges (!), etc), we need some sort of binary gate-keeping system. It would be good if we had a better one than the current glamourized journal system.
PS. I have previously said that peer reviews used for editorial gate-keeping shouldn't be published. That is because those peer reviews are "advice to the editor" to make the binary journal yes/no decision. They serve a different role than the eLife public assessment. I would be fine if these "reviews are published" assessments included three separate components:
(1) a private message to editor and authors that is designed to make the paper better [not published],
(2) a private message to the editor and authors about whether the paper should enter the literature as is (or at all) and if fixes could make it suitable to enter the literature [not published],
(3) a public assessment of the quality of the work [this is published]

2024-06-29 19:30:34 UTC

- reply

Obviously, real evaluation is not binary, but yes, the main point of peer review is for an expert without a conflict of interest to evaluate the work so that we can know what to build on as we move forward.
Science is a bunch of technologies we've developed to help us not fool ourselves (as a community). True peer review is one of the best.
As we've all talked about endlessly, the encoding by the fact that it's been published in a journal is a poor proxy for the real peer review.
We need peer review because none of us are experts in everything, but unless we want to live in ever more siloed ivory towers with no meaningful consequences, we will need to integrate our work with work from fields that we are not experts in. Thus, we need peer review to determine the quality of that work and whether it is a foundation on which to build or not.

2024-06-29 19:03:48 UTC

- reply

Unpacking: "What we need review for is to provide context for people who are not in the field."
Within a field, we can do our own "peer review". As a hippocampal expert, I can read a paper about hippocampus and do my own review. I can read between the lines and determine whether they have done the right controls. I know the literature and can tell if there are other alternative explanations that are more compatible with the rest of the literature than the speculation they've chosen. I can decide whether to build on their work or to ignore it. This is much of what we do in my lab's journal club. I assume that other experts do this to our papers in their journal clubs.
But, outside of my field, we cannot do our own "peer review". I can't read a paper on neanderthal burial rituals and determine if they've done a good job with the work. I need an expert to provide that context, to tell me if they've taken into account all the other data that exists in that literature, if they've done the controls that are expected from that field of science, if they've made any of the mistakes that often happen when using that technology.
This is particularly important when we start using science outside of the confines of its field. If we are going to decide whether to trust a vaccine or to eat horse-dewormer-paste or to change addiction policy, we need to have evaluation (review) by an appropriate scientific community of experts (peers). That process can be gate-keeping by journals or through post-evaluation discussion, but this review cannot just be some random volunteer, it needs to be an expert in the field who knows what to look for. That means it is serious work that can only be done by a very limited set of people and that takes serious time. Thus it requires some reason to take the action (usually an editor cajoling a scientist into "doing it for the community").
Peer review has also been used to help the authors make the paper better. While it is absolutely true that every paper I have ever published is better for the peer review that was done for it, it is very rare that the peer review has truly changed the meaning of the paper or the implications of the results. And there are other pathways that also provide this kind of feedback (discussions at conferences, questions at talks, post-publication commentary, papers written in response, etc.) Honestly, I think peer review does this better than any of other systems, and so I am loathe to lose it, but the cost of peer review is high and the other systems do work acceptably, especially if you are willing to take the long view on science.

2024-06-29 15:14:39 UTC

- reply

Who selects the reviewers for a PubPeer review?
The problem with PubPeer is that it is (as I understand it) a commentation system for people who have something to say about an article. But that's not what we need review for. What we need review for is to provide context for people who are not in the field.
A good review is a lot of work. It needs to be from a separate group (so as to avoid conflict of interest). It is not supposed to be self-serving (so it shouldn't be because someone feels they need to add to the discussion). Instead, a proper review is that someone asks a qualified expert to provide context and evaluation of a product.
The first key (conflict of interest) is why we can't just leave preprints unreviewed. The second is why we can't just leave it open for people who are helpful. The third is why editors ask reviewers to provide reviews, and why we don't want authors to be the ones soliciting the reviews.

2024-06-29 12:35:05 UTC

- reply

I assume you mean "post-publication" in the sense of post-preprint.
In my experience, none of the "let's let readers provide comment/review if they want" has worked because it's really rare that people want to write the deep-dive that a good review requires. I don't know whether that is because people don't want to do the work of reading things that carefully or if they don't want to do the work of writing their thoughts down or if they don't want to put themselves on the line (even an anonymous review is putting things on the line), but people don't seem to do that. Certainly people don't do that at the scale that science requires.
So, I do think we will still need "reviewers" who are requested / cajoled into doing it by editors, but I agree that there's no need for it to be pre-publication rather than post-publication.
One interesting thought is that if one could get paid for being a reviewer, people might do that. (One of my recent postdocs said that this would be a job they'd like.)

2024-06-29 11:59:54 UTC

- reply

"Multiple journals OK, as long as you list only one line on your CV" would be more work for reviewers, but if the journals are in vastly different fields, then the reviewers might be looking for different things in the paper.
If we follow some of the suggestions of people like @npub1x5a…yh0y or @npub14c5…rfvf that the future of journals is as "collections" of preprints with commentary (much like #eLife is doing now), then one might have different commentary in different fields.
Think of, for example, a neuroeconomics paper with one set of peer review / commentary in a neuroscience journal and another set in an economics journal.
On the other hand, this would likely open the floodgates of simultaneous submission to multiple journals in the same field which would be a nightmare. (How far apart would two journals have to be to be OK to publish in both?)
Personally, I'm definitely not going to do multiple journals, but I very well might do multiple preprints, particularly for specific cross-field cases (like neuroeconomics).

2024-06-28 19:08:43 UTC

- reply

So with 95 responses, the overwhelming response is that only one peer reviewed journal is allowed. But it's about 50/50 saying whether you can do multiple preprint servers.
(For completeness, 8% said multiple journals OK, and 1% said preprint or journal, not both).
#Preprints #SciencePublishing #AcademicChatter #Science

2024-06-19 10:58:55 UTC

Curious what the community thinks...
Is it OK to release the same manuscript / paper to multiple preprint servers? (Presumably because different communities look for preprints in different places.)
I was always taught that it was not kosher to publish the same paper in multiple journals (that's "double dipping" and self-plagiarism). But most people release papers to a preprint server AND publish the same paper subsequently in a peer-reviewed journal.
Note: we can assume, for the sake of discussion, that this will only be one line on a CV. (In the same way that people say "... in journal X... also available as a preprint at ...", people can say "... in journal X... also available as a preprint at ... and at ...")
Thoughts?
Please boost for discussion.
#Preprints #SciencePublishing #AcademicChatter #Science