Tim Behrens @behrenstimb Twitter profile

Last Seen Profiles

@ida_skibenes

@bokeplokalmalam

@dhawatt

@FeralKonig

@jellygrave1

@Claws_2308

@AbwDant82488

@baniazzss

@deonputraorion

@DhatWookie

@zkikro

@todokugou

@gaesolimo

@lcllnnfrl

@stw46

@FGLeaksAndInfo

@radiofeyalegria

@ugureskici

@craigharper7

@JerryWP4241

@MRen_09

@williamngriffin

@yogirlmomae

@CinemaCon

@brhH_5

@RobertMalm

@cheineze

@Aljabe123

@BinorRaja

@thalitaquino

@taekchu

@600years_

@AImodrj

@hirepangs

@stw_pdg

@big_data_kane

Tim Behrens

@behrenstimb

6 years

Twitter-sized history of neuroscience (biased by my interests).

31

728

1K

Tim Behrens

@behrenstimb

3 years

I had the privilege to be on a zoom call with Francis Collins (Outgoing NIH director) today. I asked him why NIH didn’t give full backing to open non-profit publishing. He replied this would likely lead to NIH funding cuts due to commercial journals’ lobbying power in US govt.

35

361

1K

Tim Behrens

@behrenstimb

3 years

My 7yo son imagines fancifully ambitious Lego projects, pathetically absent in any detail. He then demands that I execute them, and regularly checks on my progress without offering any constructive input. I think he is destined to be a PI.

2

50

974

Tim Behrens

@behrenstimb

3 years

Giving up neuroscience.

14

12

498

Tim Behrens

@behrenstimb

3 years

Come on brain scientists. 2022 is the year we’re going to solve it. For sure.

17

37

430

Tim Behrens

@behrenstimb

8 years

Some notes about starting a lab.

12

197

410

Tim Behrens

@behrenstimb

4 years

New service available to all. Will retweet for $9500. If you pay $2600 (non-refundable), I will advise you on whether your tweet is good enough to be retweeted, and may or may not divert you to a colleague with a lower follower count.

15

26

410

Tim Behrens

@behrenstimb

6 years

This conversation that just appeared in my inbox is one of the reasons why @eLife is so cool.

2

49

368

Tim Behrens

@behrenstimb

6 years

The recent open reviewing debate is a reminder to treat people like humans first and scientists second. Being kind has always been, and remains, more important than being right.

12

54

362

Tim Behrens

@behrenstimb

5 years

1700 people currently at #neuromatch2020 , a conference which did not exist 3 weeks ago and all of them in the comfort of their own homes. Totally unbelievable. An amazing achievement by @KordingLab @neuralreckoning @titipat_a et al.

5

68

361

Tim Behrens

@behrenstimb

2 years

First reviewed preprint published under eLife's new model. Posted to biorXiv Nov 21st. Reviewed preprint posted 2 months later. Looks like a paper. Authors can choose to answer the reviewers and have it updated. Or leave it. Here it is. A Revolution:

14

75

337

Tim Behrens

@behrenstimb

10 months

This is one of the most striking optogenetic effects I have ever seen. Stimulate the Gabaergic cells in the mrn and the animal can't stop what it is currently doing, even if aversive. Stimulate the Glutamatergic cells and it switches to something else.

A subcortical switchboard for exploratory, exploitatory, and disengaged states

To survive in evolving environments with uncertain resources, animals need to dynamically adapt their behavior and exhibit flexibility in choosing appropriate behavioral strategies, for example, to...

www.biorxiv.org

9

70

328

Tim Behrens

@behrenstimb

4 years

Tested COVID+ in an @ONS trial 6 days ago. First symptoms today - 6 days later. Wear masks people. You never know if you got it.

79

33

303

Tim Behrens

@behrenstimb

4 years

However far you progress in science you always think that there is a set of people who know more than you. Who are looking down at you. Then later you always discover they didn’t, and they weren’t. They were just as baffled as you were.

8

28

298

Tim Behrens

@behrenstimb

4 years

Exciting new initiative at @elife . In the long term we aim to separate the evaluation of science from its publication, so work is evaluated after it is published. We are starting by offering a new preprint review service. We will review preprints and publish the reviews online.

eLife - the journal

@eLife

4 years

Beginning today, we are pleased to invite authors who share their work as a preprint on bioRxiv to submit to our new @PreprintReview service

1

154

378

10

114

298

Tim Behrens

@behrenstimb

3 years

A long time ago I was sitting in Demis’ office in Russell Square. Demis said he was going to use AI to solve everything. It just seemed so ridiculous that I laughed. Who’s laughing now?

nature

@Nature

3 years

Research published in Nature reports a deep reinforcement learning architecture to design controllers for the magnetic confinement of plasmas in tokamaks

1

19

76

12

42

322

Tim Behrens

@behrenstimb

3 years

One thing about this job that never gets old is that sometimes, just occasionally, you know something about the world that nobody else knows. Happened in my group today. Still a big buzz 20 years in.

9

11

281

Tim Behrens

@behrenstimb

6 years

Wow. Couldn’t disagree more. (1) Don’t be strategic. Follow your interests.(2)Learn technical skills & ingest knowledge. You’ll write better papers and make better decisions. (3) Collaborate on lots of projects. You’ll learn loads. (4) One deep paper is worth ten shallow ones.

9

63

281

Tim Behrens

@behrenstimb

5 years

Seems to me that ML has suffered massively by having conf proc. as its major output. No space for clarity. No serious peer review. Ridiculously short turn-arounds so people work on 6-month projects not deep ideas. Keep journals sacred in neuro please. Even with their faults.

Ben Poole

@poolio

5 years

peer review in machine learning is broken #ICLR2020

12

42

280

14

41

265

Tim Behrens

@behrenstimb

4 years

Brief thread about this, particularly about Maimuna Hassan, who is featured: Muna got a place at @Newnham_College , @Cambridge_Uni . She lost it due to the A-level fiasco. She was down-graded due to her school's historic performance. I know her.

A-level results: 'My future has been set back completely'

Students share their disappointment after their predicted exam results were downgraded.

www.bbc.co.uk

12

124

265

Tim Behrens

@behrenstimb

3 years

Recommendation-letter request number 17 of 2022, but this one comes with these excellent tips for gender-balance, so I will forgive it.

9

63

265

Tim Behrens

@behrenstimb

4 years

Did anyone have any ideas since March? We have managed to get "process" stuff done (finishing off some papers etc), but where we were stuck we are still stuck, and we haven't thought of anything new for 6 months. I think you need to hang out with people for that.

16

13

264

Tim Behrens

@behrenstimb

4 years

Hey Germans - Can we borrow Angela Merkel for a few months? We need her more than you do right now. You can have her back some time around March and she'll bring home a nice trade deal with her. Will be best for everyone, I promise.

Eric Feigl-Ding

@DrEricDing

4 years

Listen to a real world leader who actually understands the coronavirus pandemic, and can even describe doubling time and can forecast model. So damn refreshing. Why can’t we have that instead of Trump??? #COVID19

627

3K

9K

11

21

258

Tim Behrens

@behrenstimb

4 years

Today was a bit rough. My fault I know, but still. Thanks to my amazing and brave lab and to friends old and new who have offered support.

10

2

246

Tim Behrens

@behrenstimb

5 years

Excited to take a new role helping the next phase of @elife .

eLife - the journal

@eLife

5 years

Today we’re delighted to welcome Timothy Behrens ( @behrenstimb ) as a new Deputy Editor

1

4

32

8

7

246

Tim Behrens

@behrenstimb

3 years

Oh dear. Look at this ridiculous attempt at summarising all the recent hippocampal state space models. Laughable. (👀 @jcrwhittington , @d_mccaffary , @BakermansJJW ) [2202.01682] How to build a cognitive map: insights from models of the hippocampal formation

How to build a cognitive map: insights from models of the...

Learning and interpreting the structure of the environment is an innate feature of biological systems, and is integral to guiding flexible behaviours for evolutionary viability. The concept of a...

arxiv.org

4

44

240

Tim Behrens

@behrenstimb

3 years

This is what I always thought science would be like.

3

53

242

Tim Behrens

@behrenstimb

5 years

If you formalise Howard Eichenbaum’s ideas about relational memory, you can predict hippocampal/entorhinal spatial responses, but also complex non-spatial things like counting cells, and why hippocampus is important for transitive inference. @jcrwhittington @TimMuller1

bioRxiv Neuroscience

@biorxiv_neursci

5 years

The Tolman-Eichenbaum Machine: Unifying space and relational memory through generalisation in the hippocampal formation #biorxiv_neursci

1

50

133

7

93

238

Tim Behrens

@behrenstimb

3 years

For your pleasure, and because it exists, here is a photo of ⁦ @nathanieldaw ⁩ having his nails painted by my daughter.

9

2

239

Tim Behrens

@behrenstimb

4 years

Complex organised behaviours can be computed by evolutionarily preprogrammed circuits outside the brain.

8

57

238

Tim Behrens

@behrenstimb

5 years

For the first year of my PhD the Z direction was defined negatively for the gradients so all of the diffusion tensors were pointing the wrong way. For the first year of analysing my 2007 data all of the choices were coded the wrong way in the behaviour but not in the task codes

2

33

229

Tim Behrens

@behrenstimb

2 years

We'd like to thank the reviewer for this excellent suggestion.

Brooklin Nash

@realBrookNash

2 years

How do you say “fuck that” in a professional setting?

1K

180

2K

9

12

229

Tim Behrens

@behrenstimb

4 years

From the amazing @nikasamborska . mPFC neurons generalise task representations across related problems. Hc neurons also have task representations, but more tied to the specific problem. Important for understanding how abstract knowledge is mapped to new situations.

bioRxiv

@biorxivpreprint

4 years

Complementary task representationsin hippocampus and prefrontal cortexfor generalising the structure of problems #bioRxiv

0

15

65

3

62

224

Tim Behrens

@behrenstimb

3 years

We started a rodent ePhys project 4 years ago. Submitted a year ago. We got reviews back requiring a huge amount of work, which made the paper better, but did not change the key results. Today we got second reviews back, requiring a new expt. How do ePhys folk build a career?

25

11

221

Tim Behrens

@behrenstimb

7 years

Yearly reminder: Everyone is trying to submit papers before Xmas and nobody wants to review them. Editors are tired and grumpy. Wait till Jan.

9

92

214

Tim Behrens

@behrenstimb

3 years

This is awesome: Grant (not mine) relies heavily on preprint. 3 out of 3 reviewers have read the preprint carefully and assessed it thoroughly. Here, at least, preprinting transformed how the reviewers read the grant. Imagine this just 5 years ago.

3

23

214

Tim Behrens

@behrenstimb

2 years

Cathy Price and I signed this book yesterday next to heroes of science past and present. I felt like a charlatan and a fan-boy. So nervous I could barely write my name, then I nearly fell off the stage. Parents and Mrs B v proud though.

13

2

200

Tim Behrens

@behrenstimb

4 years

If you want to learn linear manifold methods, or just brush up on linear algebra, @BakermansJJW has made a matlab tutorial to accompany @SaraASolla 's @worldwideneuro talk. Talk: Code: Featuring Cats and Camels in latent spaces.

3

46

198

Tim Behrens

@behrenstimb

4 years

There are really only two questions. Is the claim interesting? Does the data justify the claim? Everything else is subsidiary. Most importantly: It is their paper not yours. Review the paper in front of you. Don't ask them to write the paper you would have written.

Tara van Viegen

@TaraViegen

4 years

I got invited to review a paper for the first time🥳 Please list your best advice below to NOT be reviewer #2 .

337

98

2K

7

39

199

Tim Behrens

@behrenstimb

5 years

Yearly reminder that everyone is trying to submit papers before Christmas and nobody wants to review them. Wait till Jan if you don't want grumpy editors and 10th choice reviewers.

8

52

191

Tim Behrens

@behrenstimb

4 years

One good thing about neuralink making a bmi is that it shines a light on what's possible, and what has already been achieved, with animal research in neuroscience. I am impressed that they are being so open about this and think this will be good for neuroscience.

4

16

181

Tim Behrens

@behrenstimb

4 years

Publishing should be free. Articles should be open access. Reviewers and editors should be paid for their time. But this needs publishing to be publicly funded as a central part of the research infrastructure. This is the right destination but it is a long road to get there.

Jen Heemstra

@jenheemstra

4 years

Anyone else see a problem with journals charging $$$ to publish open access *while* paying reviewers $0 *and* saying they need to recruit more reviewers from underrepresented groups?

137

654

6K

4

27

181

Tim Behrens

@behrenstimb

4 years

Theorist visits an experimental lab.

1

12

177

Tim Behrens

@behrenstimb

5 years

This was my favourite thing at Cosyne this year. The further you get into it, the deeper it goes.

High-dimensional geometry of population responses in visual cortex

Nature - Analysis of the encoding of natural images by very large populations of neurons in the visual cortex of awake mice characterizes the high dimensional geometry of the neural responses.

www.nature.com

7

50

170

Tim Behrens

@behrenstimb

2 years

I confirm I have raised enough support to run for @elife EiC, and believe I am the only candidate with a mandate to rectify @mbeisen 's reckless policies. However, I have come to the conclusion that this is not the right time.

10

2

171

Tim Behrens

@behrenstimb

2 years

By associating journal name with quality, the current system distorts science and slows down progress/careers. We need a system where name signifies quality of *assessment* not the science. Major changes at @elife will implement such a system:

7

35

160

Tim Behrens

@behrenstimb

5 years

Really clear and thoughtful review of an amazing decade of thinking by @yael_niv

Learning task-state representations

Nature Neuroscience - When crossing the street, you can ignore the color of oncoming cars, but for hailing a taxi color is important. How do we learn what to represent neurally for each task? Here,...

www.nature.com

2

38

163

Tim Behrens

@behrenstimb

3 years

Rejected from @biorxiv_neursci . Bloody Glamour journals.

12

4

157

Tim Behrens

@behrenstimb

4 years

Scientists discover a new technique.

1

14

160

Tim Behrens

@behrenstimb

3 years

Always nice when this happens @elife .

1

160

Tim Behrens

@behrenstimb

5 years

Gyorgi Buzsaki wrote a comment on our paper. This must be how Royalists feel when they meet the queen.

3

24

155

Tim Behrens

@behrenstimb

6 years

This is a really beautiful paper. It is exactly what theory should be. A clear concise formal argument to explain a wealth of seemingly disparate data. Very pretty @nathanieldaw @marcelomattar

Nature Neuroscience

@NatureNeuro

6 years

A normative theory predicting which memories should be accessed at each moment to optimize future decisions

3

112

209

1

60

152

Tim Behrens

@behrenstimb

3 years

Cannot begin to say how excited I am that @saadjbabdi has been made full professor at Oxford. Nobody has ever deserved it more.

15

3

155

Tim Behrens

@behrenstimb

2 years

If you are writing appeal letters, I think that broadly there are only 3 successful successful arguments: 1. The reviewer is factually wrong 2. The reviewer has not understood 3. You have new data proving the reviewer is wrong. "We disagree with the reviewer" never works.

18

16

151

Tim Behrens

@behrenstimb

6 years

Yearly reminder that everyone is trying to submit papers before Christmas and nobody wants to review them. Wait till Jan if you don't want grumpy editors and 10th choice reviewers.

5

32

152

Tim Behrens

@behrenstimb

4 years

!!!!!!!!

eLife - the journal

@eLife

4 years

From today, code can be an integral part of a paper published in eLife. Announcing the Executable Research Article, a new open format for research papers where code can be run directly in the browser

9

308

746

2

31

149

Tim Behrens

@behrenstimb

5 years

No model selection technique works well enough to test a scientific claim. Design a better experiment.

Daniela Witten

@daniela_witten

5 years

Please quote tweet this with your most controversial opinion about model selection. #rstat

7

0

33

5

27

150

Tim Behrens

@behrenstimb

7 years

Apparently you shouldn't use the word "dickhead" in reviews, even when referring to yourself.

9

52

143

Tim Behrens

@behrenstimb

2 years

Final version of @nikasamborska 's awesome generalisation paper now online: Data: Code: Excitingly: last author Thomas Akam has just won a @wellcometrust fellowship to start his own lab - look out for jobs!

GitHub - veronikasamborska1994/notebooks_paper

Contribute to veronikasamborska1994/notebooks_paper development by creating an account on GitHub.

github.com

6

36

145

Tim Behrens

@behrenstimb

4 years

I am starting to think I should sign up for next year’s @neuromatch academy as a student.

Miguel Núñez 🇲🇽

@NunezKant

4 years

I would say that today I had the best GLMs explanation ever, In fact I already had did that exercise about spike triggered average somewhere else but in this case the inner workings of the maths underlying the GLM were exceptionally explained Every day you amaze me @neuromatch

4

11

85

9

8

146

Tim Behrens

@behrenstimb

4 years

I have an OFC joke but I can't take credit for it I have another OFC joke but I don't know what state it's in I have a Hippocampus joke but I can't remember it I have a Bayes joke but I can't integrate it I'll stop now.

Ida Momennejad

@criticalneuro

4 years

I have a reinforcement learning joke, but not sure it's rewarding.

31

75

848

15

9

145

Tim Behrens

@behrenstimb

7 years

Dear @nature @ih_C_hi . We know Optogenetics is cool but this is a poor reflection of neuroscience in 2017.

8

98

142

Tim Behrens

@behrenstimb

4 years

Similarly, the discovery of place cells was met with heavy resistance from reviewers and was published in Brain Research (IF=3). 43 years later it won John O’Keefe the Nobel Prize. People are often not ready for something new.

Oded Rechavi

@OdedRechavi

4 years

The discovery of ubiquitin-mediated proteolysis was published in the journal BBRC Biochemical and Biophysical Research Communications IF=2.95 after it was rejected from the fancy journals. 26 years later it won Ciechanover and Herschko the Nobel Prize

16

209

1K

4

22

144

Tim Behrens

@behrenstimb

3 years

Quick thread contextualising this spectacular paper. Briefly, I think it is an extraordinarily powerful example hippocampus creating a latent state space.

Edward Nieh

@enieh36

3 years

It's finally out! It was a slugfest, but I had a fantastic time working with @MSchottdorf @sueannkoay @lucasmpinto @JeffOnTheBrain @brody_lab .

9

134

506

1

22

143

Tim Behrens

@behrenstimb

4 years

Rather disappointed to co-author my first paper with @nathanieldaw today, after 15 years of bitter animosity.

6

1

137

Tim Behrens

@behrenstimb

4 years

This is amazing by @doristsao and team, again.

A map of object space in primate inferotemporal cortex

Nature - Primate inferotemporal cortex contains a coarse map of object space consisting of four networks, identified using functional imaging, electrophysiology and deep networks.

www.nature.com

2

45

136

Tim Behrens

@behrenstimb

4 years

Lab zoom - who owns which brain areas, and the most exciting fights.

43

5

134

Tim Behrens

@behrenstimb

3 years

Developing quite an immunity collection. Vaccines:2xAz, 1xModerna. Infections:1xWT, 1xAlpha, 1xOmicron (probably). Just need Delta to complete the set.

24

2

137

Tim Behrens

@behrenstimb

5 years

Early contender for sysneuro paper of the year.

Uncovering the Neural Basis for Hypothetical Thinking in Rats | HHMI

As a rat approaches a maze juncture, neural activity rapidly cycles between two sets of cells. The discovery suggests the rodents are envisioning turning left or right. A similar phenomenon may...

www.hhmi.org

2

31

129

Tim Behrens

@behrenstimb

5 years

Official rankings of NYC neuroscience institutes by important academic criteria: Amount of glass (1) Columbia (2) Princeton (3) NYU Medicine (4) Penn (5) NYU downtown

3

11

134

Tim Behrens

@behrenstimb

3 years

Not sure why anyone would want to know my views on life, but if you are bored you can find some of them here :)

Neuron

@NeuroCellPress

3 years

Read the final part of our Special Issues series. We have #NeuroViews on diversity, equity, and inclusion & 7 Q&As from inspiring neuroscientists

0

37

101

7

19

130

Tim Behrens

@behrenstimb

5 years

“If a child grows up to be a scientist he finds that he is paid to play all day at the most exciting game ever devised by mankind.” or she. Otherwise, perfect.

In memoriam: Murray Gell-Mann

Murray Gell-Mann, a Nobel laureate who revealed symmetry and order in the world of subatomic particles and leveled his genius at complex mysteries of life and mind, died peacefully May 24, 2019. He...

www.santafe.edu

1

28

129

Tim Behrens

@behrenstimb

6 years

Two beautiful papers that should get everyone thinking about how tasks, goals and policies are represented in our brains. Very exciting.

0

44

128

Tim Behrens

@behrenstimb

4 years

6yo: If I have COVID can I play Nintendo all day? Me: Yes 6yo: Ok, if you catch it I will let you play on twitter all day.

3

1

127

Tim Behrens

@behrenstimb

3 years

Super paper

0

29

126

Tim Behrens

@behrenstimb

1 year

Some exciting news. The UCL part of my lab is moving to @SWC_neuro . The Oxford part is staying at @OxfordWIN , and I hope to keep collaborating closely with @WCHN_UCL . I am really looking forward to combining the circuit busting prowess at @SWC_neuro with ideas from cogneuro.

SWC

@SWC_Neuro

1 year

We are delighted to announce that Tim Behrens ( @behrenstimb ) has joined SWC as a Group Leader. Find out more about his research on how brains learn and represent knowledge about the world to enable flexible behaviours:

0

4

59

5

9

126

Tim Behrens

@behrenstimb

6 years

2015-2020: This neural network looks just like my neurons. 2020: Jesus that’s a lot of neurons.

5

14

119

Tim Behrens

@behrenstimb

6 years

Big step towards a fun place. Articles with embedded data and code that let you really interact with the data.

3

46

122

Tim Behrens

@behrenstimb

5 years

New set of policies for @elife papers. We hope this will make publishing in @elife as joyful during a pandemic as it is in usual times. In particular we hope it will help early career researchers who no longer have access to labs. Thread 1/n

eLife - the journal

@eLife

5 years

We're making changes to our peer review policies in response to the impact of #COVID19 , including curtailing reviewers' requests for additional experiments and extending our ’scoop protection’

0

38

113

4

45

121

Tim Behrens

@behrenstimb

3 years

For those who don't know, Veronika is a neuroscientist at Oxford studying generalisation and inference in frontal cortex. If the neuro community can help it would be great.

1

58

117

Tim Behrens

@behrenstimb

3 years

I am totally excited that @CosyneMeeting is going to be in Montreal. But I wish that, just once, it would be honest and use a mouse brain on the poster instead of a human brain. @justsaysinmice

xaq

@xaqlab

3 years

Looking forward to #cosyne23 !

4

19

182

3

7

120

Tim Behrens

@behrenstimb

4 years

Good luck tomorrow, America. Fingers and toes crossed for you.

1

2

118

Tim Behrens

@behrenstimb

4 years

Very grateful to the British weather for taking us to every climate zone when we haven’t been able to travel this summer. Today we appear to be in the middle of the Atlantic Ocean.

2

4

117

Tim Behrens

@behrenstimb

5 years

Our latest from @BaramAlon @TimMuller1 @hamed_nili @mona_garvert . Using a "task remapping" paradigm, we show reinforcement learning problems are abstracted, and these abstract representations can be generalised to new problems.

bioRxiv Neuroscience

@biorxiv_neursci

5 years

Entorhinal and ventromedial prefrontal cortices abstract and generalise the structure of reinforcement learning problems #biorxiv_neursci

0

12

35

2

47

114

Tim Behrens

@behrenstimb

4 years

In case anyone is interested what it looks like when you do something stupid on twitter. Hope there isn't a second wave.

13

2

113

Tim Behrens

@behrenstimb

8 months

Data from the first year of our new approach to publishing!! In brief, we are still seeing the same quantity and caliber of papers. But the process feels much more constructive. For Editors, and for authors. Data: Editorial:

eLife’s New Model: One year on

Explore key milestones, achievements and findings from the first year of eLife’s new model.

elifesciences.org

eLife - the journal

@eLife

8 months

Thousands of authors have put their trust in our vision for publishing. Today, we look back at the first year of our new model in numbers. 1/3

2

18

49

3

39

110

Tim Behrens

@behrenstimb

4 years

This is really really cool. Matthias showed me it on the @neuromatch meet. Can decide gaze position accurately without an eye tracker, straight from MRI signal.

Matthias Nau

@NauMatt

4 years

If you are at #VSS2020 & like #fMRI , #eyetracking & #CNNs , drop by my poster! @CYHSM & I developed #DeepMReye , a CNN-framework to decode gaze position from the MR-signal of the eyeballs. Post-hoc in held-out data! No camera! Even eyes-closed! @doellerlab @KISNeuro @MPI_CBS 1/2

11

63

202

4

19

114

Tim Behrens

@behrenstimb

2 years

I know I am biased but I think this is the coolest thing in the history of the world. The neural representation completely reveals the algorithm the mice uses to solve a complex abstract inference, and store a long probabilistic sequence of actions.

Mohamady El-Gaby

@GabyMohamady

2 years

Want to know how mice do clever inferences the first time they see a new task? I’ll be presenting our latest findings with @behrenstimb on Frontal mechanisms of abstraction tomorrow, 3pm BST as part of the neuromatch conference.

1

8

82

2

15

111

Tim Behrens

@behrenstimb

4 years

Code for Tolman Eichenbaum Machine now online. Can reproduce all modelling results in paper (up to a random seed). Email all of your queries and complaints to @jcrwhittington .

GitHub - djcrw/generalising-structural-knowledge

Contribute to djcrw/generalising-structural-knowledge development by creating an account on GitHub.

github.com

Tim Behrens

@behrenstimb

5 years

If you formalise Howard Eichenbaum’s ideas about relational memory, you can predict hippocampal/entorhinal spatial responses, but also complex non-spatial things like counting cells, and why hippocampus is important for transitive inference. @jcrwhittington @TimMuller1

7

93

238

4

17

112

Tim Behrens

@behrenstimb

8 years

Tip from an editor to people trying to submit papers before Xmas. Everyone's doing it, and nobody wants to review them. Wait till Jan.

11

106

111

Tim Behrens

@behrenstimb

4 years

Just got reviews for our own paper from @PreprintReview . Fair reviews but still scary to think they'll be published. *But* @elife decision was "reject with option to resubmit", which will take time. @PreprintReview means it is already published and evaluated! This is the future.

0

12

111

Tim Behrens

@behrenstimb

6 years

Seriously? seems like every two-bit loser with a pair of lungs has an opinion on what a cognitive map is these days.

bioRxiv Neuroscience

@biorxiv_neursci

6 years

What is a cognitive map? Organising knowledge for flexible behaviour #biorxiv_neursci

0

32

42

9

31

107

Tim Behrens

@behrenstimb

3 years

Here are our submissions for Cosyne 2022. Broad theme: Task Abstractions/Coordinate Systems and how to use them. You'll have to come to Lisbon to find out which were accepted.

3

18

107

Tim Behrens

@behrenstimb

6 years

Jesus

Nature Portfolio

@NaturePortfolio

6 years

Research published in Nature describes an artificial neural network made out of DNA that can solve a classic machine learning problem: correctly identifying handwritten numbers. The work is a step towards programming AI into synthetic biomolecular circuits

10

334

482

0

40

105

Tim Behrens

@behrenstimb

7 years

3-week late reviewer is having a pointless 4-hour twitter argument. Should the editor:

Politely ignore

84

Send angry email

22

RT to show I'm watching

149

Do passive agressive poll

456

6

30

100

Tim Behrens

@behrenstimb

4 years

Can @BBCRadio4 please stop giving airtime to anti-vaccination people please. Particularly with their uninformed scare stories. It is total madness. It can only lead to bad outcomes. And can academics please be less “academic” in putting them right.

6

15

100

Tim Behrens

@behrenstimb

6 years

The atmosphere was really amazing on the march today. I’ve never known anything like it. I hope it gives MPs some strength to vote with their consciences over the next few weeks.

1

6

99

Tim Behrens

@behrenstimb

2 years

Figuring out how the brain works seems hard enough to me. I'm happy to let someone else solve AI.

3

5

103

Tim Behrens

@behrenstimb

2 years

Thanks very much inviting me to speak at your US conference. I’m afraid I can’t make it. Thanks very much for offering me a postdoc position in The U.K. but I’ve been offered twice the salary in US.

3

6

103

Tim Behrens

@behrenstimb

4 years

*These authors contributed equally to this manuscript.

4

15

103