US Forbids Open Access Publishing?

There's a proposed law in the US to prohibit researchers paid with public money to, well, make the results available to the public. They would not be allowed to publish in Open Access journals or to submit to public-access repositories. And not surprisingly it is slimeball companies like Elsevier (you know, the charming guys who created fake medical journals so pharmaceutical companies could push made-up data in support for their products) and Wiley Publishing that are behind it.

I've said it before, but I will no longer publish anywhere that is not Open Access. OK, if you look at my publication record that's not exactly a threat of earth-shaking significance. But I will also no longer review papers for closed journals. I do review quite a lot of papers; it's fun and it's a good way to stay abreast of a wider range of subjects. And publishers depend critically on a good supply of reviewers to do their (unpaid, unacknowledged) work.

I will — I already do, to some extent — make a point of using Open Access sources for citations and other resources when I can. I've said before that there's two kinds of citations in your papers. A few are central to your own work, and pretty much unavoidable. But most papers are really about establishing background for your work and give references to general knowledge, and there you often have any number of papers to choose from. The same group may have half a dozen papers that all cover the point you want to make, there may be several groups all working on the same thing, and any of them would be fine as a general reference.

 
I will make a special point of not dealing with Elsevier in any way, shape or form. Which leaves me with one issue: I'm a member of the Japanese Neural Network Society. I need to be a member to be able to attend and publish in their conferences and meetings. They have a Japanese journal of their own, but also co-publish an English language journal together with the European and International societies. Published by, you guessed it, Elsevier.

Resigning from the society is unproductive, and feels like overkill. I'm not dealing with Elseview when I'm going to a conference or something after all. As I'm not Japanese¹ and just a simple post-doc I have zero clout in the society, protesting a long-standing publishing arrangement is futile. What I will do is simply ignore the English-language journal. Not publish, not review, and, where feasible and honest, not reference. Should not be too difficult as I have yet to do either.

Remember, while this suggested law is recent and US-centric, the instigators — the for-profit journal publishers — have been fighting to stay gatekeepers of research across the globe for years. It is a very profitable business, and they have shown there is little they would not do to stay in it no matter how much it hurts science, the scientists who they depend on, or the public that pays for it all.

#1 It's a question of language. Your ability to convince others about a controversial position is tightly coupled to your ability to make a strong rhetorical case. As my Japanese is barely usable I am unable to make good, coherent arguments, fully understand voiced objections, or formulate convincing answers.

Badger Problems?

The title in this recent paper in PLoS One says it all, really: Effectiveness of Biosecurity Measures in Preventing Badger Visits to Farm Buildings. What is a paper on badger control doing in a highly respected science publication?

For one thing, it shows us an important fact about science: Science is a method. It is not about what you study, but about how you study it. Preventing badger incursions is legitimate science when you go about finding out in the proper way.

What is that proper way? There's many descriptions out there, focusing on philosophical or practical aspects. But the gist really is that you find answers to questions without getting fooled — fooled by bad data, fooled by badly posed questions or fooled by your own biases and expectations. It boils down to making sure that your results are real, not an artefact of noisy data or influenced by what you wish were true.

This paper is a well-done bit of science, and it also addresses a real problem. Badgers visit farms, presumably to find stuff to eat, and when they do they can infect the cattle with a version of TBC. Badgers and cows don't come into direct contact very often, so the disease probably spreads indirectly, through badger faeces and urine. Stopping badger visits seems like a good idea.

But farms are big and badgers are small. Stopping them altogether is difficult and expensive. It would be great if you could use simple ways to just keep badgers out of the cattle feed storage, where the risk of spreading infection is greatest. But you also want to know what happens if you do that. Say that you stop badgers from getting into the feed but the badgers simply spend more time where the cows sleep instead, then you may have gained nothing.

 
The authors set up surveillance cameras on a number of farms. For a whole year they simply recorded the number of badger visits to various areas on the farms. Then they picked farms with badger visits (not all of them had a badger problem) and divided them into three groups: One got badger barriers around the cattle feed area; one around the sleeping area and one got both. Spend another year recording (and analysing all those those hours of recording must have been painful for some poor graduate student) then analyse the results.

What did they find? All barriers were effective in reducing badger visits, and barriers around one area reduced, not increased, visits to other areas as well. Barriers around both areas were most effective, but the feed area barriers were almost as effective.

Think of what they did to get to that result. They decided at the outset what question to ask and what would be an acceptable kind of answer. They first just recorded visits so that they'd know the baseline of visits, and they spent a whole year doing so, since the badger frequency likely changes over the seasons. Then they changed things systematically — add barriers to one, other or both areas, covering all possibilities — and recorded for another year.

And they were careful to set precise criteria for the recording and analysis beforehand — what does and does not count as badger visit, what to do with the data if a camera doesn't work part of the night, if a barrier was down, that sort of thing — so that their own expectations wouldn't influence the analysis. They were careful to use statistics to find out how likely the changes they saw were real and not just due to chance.

Whether you're looking for badgers or bosons, it's this painstaking attention to remove errors, uncertainty and biases that makes it science.

PhD at OIST on Okinawa

OIST — Okinawa Institute of Science and Technology — is looking for PhD candidates for next year.

Briefly, OIST is a new, graduate-level research institute on Okinawa, Japan. The labs range from neuroscience to math to physics to chemistry, but the work is very interdisciplinary. Up until now they have been "just" a research facility as they've been building up the organization, but from this year they are accredited as a graduate school.

They aim for a very international research and graduate student body, and all teaching and interaction is in English. About half of the staff is non-Japanese and I believe the aim is to have the same kind of proportion among students as well.

I was a tutor at the OCNC summer course on computational neuroscience last year. The facilities are very good, and there's a wide variety of research going on. The summer school itself is very high quality.

The one possible drawback with OIST is that the location is decidedly rural. Even grocery shopping is a chore without a car. On the other hand, it gives you plenty of uninterrupted time, and the location is excellent if you happen to be into diving or snorkelling.

If I were a budding graduate student, I'd grab this chance without second thoughts.

Patch clamp microscope at OIST. I was going to add a scenic picture here, but in all honesty, you're more likely to stare at something like this all day long than at any tropical island view.

Badger Problems?

The title in this recent paper in PLoS One says it all, really: Effectiveness of Biosecurity Measures in Preventing Badger Visits to Farm Buildings. What is a paper on badger control doing in a highly respected science publication?

For one thing, it shows us an important fact about science: Science is a method. It is not about what you study, but about how you study it. Preventing badger incursions is legitimate science when you go about finding out in the proper way.

What is that proper way? There's many descriptions out there, focusing on philosophical or practical aspects. But the gist really is that you find answers to questions without getting fooled — fooled by bad data, fooled by badly posed questions or fooled by your own biases and expectations. It boils down to making sure that your results are real, not an artifact of noisy data or influenced by what you wish were true.

This paper is a well-done bit of science, and it also addresses a real problem. Badgers visit farms, presumably to find stuff to eat, and when they do they can infect the cattle with a version of TBC. Badgers and cows don't come into direct contact very often, so the disease probably spreads indirectly, through badger faeces and urine. Stopping badger visits seems like a good idea.

But farms are big and badgers are small. Stopping them altogether is difficult and expensive. It would be great if you could use simple ways to just keep badgers out of the cattle feed storage, where the risk of spreading infection is greatest. But you also want to know what happens if you do that. Say that you stop badgers from getting into the feed but the badgers simply spend more time where the cows sleep instead, then you may have gained nothing.

 
The authors set up surveillance cameras on a number of farms. For a whole year they simply recorded the number of badger visits to various areas on the farms. Then they picked farms with badger visits (not all of them had a badger problem) and divided them into three groups: One got badger barriers around the cattle feed area; one around the sleeping area and one got both. Spend another year recording (and analysing all those those hours of recording must have been painful for some poor graduate student) then analyze the results.

What did they find? All barriers were effective in reducing badger visits, and barriers around one area reduced, not increased, visits to other areas as well. Barriers around both areas were most effective, but the feed area barriers were almost as effective.

Think of what they did to get to that result. They decided at the outset what question to ask and what would be an acceptable kind of answer. They first just recorded visits so that they'd know the baseline of visits, and they spent a whole year doing so, since the badger frequency likely changes over the seasons. Then they changed things systematically — add barriers to one, other or both areas, covering all possibilities — and recorded for another year.

And they were careful to set precise criteria for the recording and analysis beforehand — what does and does not count as badger visit, what to do with the data if a camera doesn't work part of the night, if a barrier was down, that sort of thing — so that their own expectations wouldn't influence the analysis. They were careful to use statistics to find out how likely the changes they saw were real and not just due to chance.

Whether you're looking for badgers or bosons, it's this painstaking attention to remove errors, uncertainty and biases that makes it science.

Data collection and analysis app, anyone?

Here's a question for people doing data analysis and programming. I'm looking for a tool that I don't know if it exists:

I often find myself collecting data over time; temperature data, my weight, baking results, lots of stuff like that. I want to be able to very quickly, simply, add data on a daily or hourly basis and do my own exploratory analysis and visualisation. Normally I use a spreadsheet, or hack together a small script to deal with the data, but neither is very convenient.

• A spreadsheet lets you enter data as it comes in, but both data entry and analysis is clumsy and rudimentary, and you soon hit the wall in what you can do with it. Try to write a spreadsheet that correlates your data with the day of the week, for instance.


• Octave, R and tools like that are very powerful. But they're not really geared for this kind of simple daily data entry and presentation.They're really about analysing fixed data sets and don't do interactive data collection very well.


• One-off tools in Ruby or Python will do what I want of course, and in practice it takes less effort than doing this kind of interactive thing in Octave and the like. But it feels like I'm reinventing the wheel every single time.

I'm really looking for a tool somewhere between a completely open-ended scripting environment and a restrictive tool like as spreadsheet; Octave or R but geared towards interactive, daily data collection rather than extensive analysis of fixed data sets.

Is there such a thing?

If not, it may be time to start thinking about creating it. A spreadsheet-like, but more task-specific, frontend, with a good way to enter new data and a real language to do your data analysis. Bonus for being able to generate a matching data entry component for Android phones (can't sideload apps on iPhone).

I'm crossposting this to Google+, and you can also reach me through email as well of course.

Real Pain, Social Pain

We often talk about emotional trouble as painful. We feel hurt by rejection, we smart from hurtful remarks, We get burned by a bad relationship, our hearts ache for company and so on. There's lots of similar expressions in other languages too; bitterness can be expressed as a form of pain in Japanese (苦痛) and your ears will hurt (耳が痛い) from hearing a painful truth. In Swedish, too, rejection and other negative emotions are painful, and experiencing others misfortune can be heart-cutting (hjärtskärande).

Emotional distress as pain is a good, productive metaphor. But — what if it's more than a metaphor? Our experience of pain is a function of our brains after all, just like emotions are. Bodily pain starts with receptors on our skin and elsewhere, but the experience of painfulness definitely happens in the brain itself.

Amputees can experience phantom pain, where the brain is led to believe there's pain in a body part that no longer exists. On the other hand, many pain relievers like morphine or codeine act on the brain pain centers rather than at the source of the pain; I've heard one person describe the effect of a similar drug as "I knew it still hurt a lot; I just no longer cared."

 

Do Pain Relievers Help with Social Pain?

Now, if emotional pain is real pain — if, in other words, "painful" emotions actually use some of the same circuits as physical pain in the brain — then central nervous system pain relievers, such as acetaminophen — commonly known as paracetamol — should work for emotional pain as well. And this is what a group led by Naomi Eisenberger set out to test recently. They published a paper, Acetaminophen reduces social pain: behavioral and neural evidence about this last year.

Unfortunately, the paper is heavily paywalled and I can't get it from the publisher. By current standards of science journalism we'd be going well above and beyond our duty simply by reading the abstract. But fortunately the authors have put up the paper on their own website: you can download the PDF right here¹. And there's another, earlier paper from Eisenberger, Why rejection hurts: a common neural alarm system for physical and social pain², that lays out a lot of the evidence for a common mechanism between physical and social pain.


They recruited two groups of participants — 62 university undergraduates in total. One group took paracetamol twice a day for three weeks, and the other one took a sugar pill, or placebo. The participants rated the amount of social pain they experienced every day. Social pain dropped significantly³ over time among those who took the pain reliever, while the people with the placebo showed no change. The pain reliever seems to lessen the pain of bad social events.

They also did a brain scanning experiment with two smaller groups, 25 people in total. The groups got either paracetamol or a placebo for three weeks. Then they got to play a simple computer ball-tossing game while lying in an fMRI scanner. They thought they played with two other people, but the game was really pre-programmed. The computer "players" gradually ignored the player and refused to toss their ball to them, making them feel rejected and left out.

They found that those who had taken paracetamol for three weeks had much less activity in two brain areas (the anterior insula and anterior cingulate) that we know are involved in the emotional aspect of pain. But there was no difference between the two groups in how painful that ball-game rejection was to them.


So it does seem that long-term doses — note that the effect took a couple of weeks to appear — of some pain relievers really can lessen the pain of social rejection. But the effect is not big, and it doesn't seem consistent. So don't go eat paracetamol on a daily basis to feel socially better — acetaminophen is not good for your liver, and especially so if you also like to drink alcohol.

fMRI scanner. That's me lying there getting ready for a scan. I wasn't ill or anything; I just volunteered as participant in an experiment. It was a fun experience, though difficult to stay awake for the entire experiment. And as a bonus you got confirmation that there's nothing obviously wrong with your brain.

…But There's More!

As it happens, another group led by Tor Wager did a similar experiment just this year (it's Open Access; anyone can read it). They recruited a group of people that had recently been dumped by their partners and stuck them in an fMRI scanner to find out what areas are involved with social rejection. They ran two sets of scans, one to find areas for social pain, and one for physical pain.

To test social pain they showed the volunteers either a picture of their ex-partner and asked them to think about their rejection; or a picture of a friend and asked them to recall a pleasant experience they've had with them. This was repeated multiple times while their brain activity was scanned. This way you can compare the brain activity with and without the bad experience, and the areas that are active only for the bad experience are probably connected to their rejection in some way.

They did the same kind of thing for physical pain: the volunteers either got burned on the arm⁴, or just pleasantly warmed on the same spot. Again, you look for differences between the painful and the non-painful heating in the brain scans. That should show you what brain areas that react specifically to physical pain, rather than to heat or things touching your arm and so on.

When they compared the two sets of differences, they found that the brain areas that deal with the emotional aspects of pain — the "feeling bad about it" — are activated by both emotional and physical pain. That's the same areas that Eisenbergers group found, and it's exactly what we'd expect. But they also found common activity in areas that deal specifically with physical pain. The emotional rejection activates areas that normally only react to bodily injury, in other words. This is surprising, and previous experiments have not found this.


One major reason, Wager's group notes, may be the level of pain. In Eisenbergers fMRI experiment, people played a simple computer game with strangers who weren't being fair to them. Not nice, but not exactly a major life crisis either.

In this experiment, on the other hand, people that have just been dumped get the picture of their traitorous ex-partner shoved into their face, and are asked to please really think through the whole sordid series of events that ended with them being thrown on the curb like yesterday's garbage. It's a whole different world of hurt, and I can imagine it took a bit of explanation to get this approved by the ethics committee.

So it may simply be that social rejection needs to be strong to actually qualify as pain. Our pain centers don't light up for every touch either; they need a minimum level of hurt to react at all. This could explain the puzzling result from Eisenbergers group, where the pain reliever seemed to have an effect for the students that reported daily social pain, but not when students were scanned. We probably encounter much worse social experiences in our daily lives than the computer ball-game they used for their fMRI scan. The pain reliever would lessen the impact of strong, but not weak, social rejection, just like it has an effect on a real skin bruise or cut but doesn't numb us to touch or slight discomfort.

…As This Is Getting Too Long Already…

The takeaway message is, I think, that the difference between our mental experiences and the physical reality is quite blurred in our brains. Paper cut or hurtful word — by the time it reaches the brain it's all just nerve inputs. There is nothing intrinsically more painful about the signals coming from your skin than from your ears. The difference is only in how our brains treat those signals.

Evolution is ultimately pragmatic. If it is useful to treat strong social rejection as physical pain then it will. It doesn't matter if it doesn't make sense from a design point of view, if it makes for a messy, untidy system, or if it will cause unintended side effects and problems for some distant descendant. The brain is full of opportunistic shortcuts, multiple-use mechanisms and exaptations which makes for a very interesting task trying to untangle it all.

--

#1 Are they breaking copyright by making their paper available like this? Maybe, and maybe not. A lot of journals do allow authors to make "draft" versions available; the difference to the published version is usually little more than the addition of magazine logos and page numbers. And depending on the legal residence of the researcher and of the journal, and on the exact wording of the contract, the researchers may retain the right — explicitly or though fair-use provisions — to disseminate their own paper.

On a more practical level, any journal that tries to sue its (unpaid, and frequently paying) contributors for passing out their own work is probably going to find themselves in a bad public relations debacle, with far worse consequences than the possibility of having lost fifty or a hundred dollars in revenue.

#2 If you're not familiar with the research world, you may not know why Dr. Eisenberger is the last author in the current paper, but the first in this earlier one. Very simplified, the first author is typically the one who did most of the actual research. The last author is their supervisor, or research leader or PI (principal investigator). They may have done parts of the actual work, but more likely provided guidance, original ideas, money and other resources.

Dr.Eisenberger, we can infer, probably did this earlier paper as a post-doc in Dr. Lieberman's lab, then managed to secure funding for her own lab, where she apparently continues her line of research but now as a leader of a group of young researchers.

#3 That is the science meaning of "unlikely to be due to chance", not the everyday meaning of "sort-of important". If you think about it, though, the meanings do overlap quite a bit.

#4 There's ways of using low heat to create intense burning sensation without any actual damage. Still, pain is not something you'd use lightly in experiments.

Research PublicationA Modest Proposal

I love doing science. But some things I love doing less than others. Rewriting papers is one of them. Editing and resubmitting a paper is to research what a wisdom tooth extraction is to a long summer weekend; all things considered you'd really rather be doing something else.

It takes a lot of time to write a long-form paper. The text may go through several revisions over the course of months even before the initial submission, and be picked-over several times by all the authors. The total time we spend may easily be a month or two. Extensive edits or a resubmission can double that time. And a lot of this work is almost invisible; we're debating commas, or precise wordings, or the order of arguments for a minor point in the text.

But very few people will actually read your paper in such detail. Most people who see it will just browse; we all "read" — that is, quickly check the summary and figures — a lot of papers, but we focus in detail only on a few. Some estimate that the average number of serious readers of a paper is around 5. And this probably follows a power-law distribution, where a small number of papers get many hundreds or thousand of readers while most papers get almost none. If your paper isn't in a top-tier journal you can probably assume your serious readership is 0-5 people.

We spend a months worth of work or more on tedious polishing. Meanwhile, almost all of our readers will simply skim the abstract, check a summary of results, look through the bibliography and then move on. Only a very few people — and perhaps nobody — will actually want to know about our work in detail.

 

So perhaps we are all spending our time on the wrong thing. I suggest we stop publishing painstakingly polished 20 or 30-page masterpieces. Instead we publish just a 2-3 page text with an abstract, a to-the-point summary of methods and results, and the bibliography. That will satisfy the vast majority of our readers, and will in fact make it easier for them to find what they want.

Then we meet by video-conferencing with those few who want to know all the details. If we save a month of work by not writing the long-form paper, and a one-hour conference takes a total of three hours with preparation and setup, then we could meet with fifty separate groups and still save valuable time. More likely, as we saw above, only a few people would ever want to discuss the details with us, saving us most of that month of project time. And those that want the details will get something better than a paper: they get the undivided attention of the researcher that did the actual work, and get precise answers to their specific questions.

We record the sessions and put them online. In the near future we'll have automatic transcription of each session as well. That will save the details for posterity and will satisfy most people looking for details, so only those with new questions and novel insights will want an in-person discussion. The just-the-facts summary and the accumulated, searchable discussions will hold far more detail, reasoning and justification of the work than any static paper could ever be able to.


Ok, so perhaps the idea isn't perfect. But it sure looks good to me while I'm sitting here editing a paper…

That Science Communication Thing

We scientists mostly suck at communicating with the public. Many of us don't even try to do any real outreach to the non-science public. We communicate our results by publishing papers or giving conference talks squarely aimed at other scientists.

This is not great. People — like Christie Wilcox here — have been calling for scientists to open up and share directly with the public. It's a nice idea. It is important that science gets reported, and it's important that people from all walks of life participate in public discourse.

Of course it's not quite that easy. Steven Hamblin brings up some important good reasons why scientists aren't communicating directly with the public, and why doing so would not have such large effect. He focuses on the skills needed for good communication. Doing it well is hard, and many, even most scientists don't have those skills.

I'd like to offer another reason: good communication takes a lot of time. This is time we in general do not have. There is no time set aside for science communication; few or no writing classes offered for faculty specifically for communicating science to the general public; no support or funds for dealing with any ethical and other issues that may arise. The scientists that do communicate directly do so on their own free time, as a hobby and elect to do it instead of watching TV, talking with their family, playing an instrument, catching up on sleep or whatever you do for fun. It is laudable that people like Scicurious does this, of course, but it is strictly voluntary.

 
And it takes real time to do it well. I only post about 2-3 times a week on my blog here, mostly not science-related, and the rate drops whenever I get busy with work. A simple off-the-top-of-my-head opinion post like this one takes me about an hour to put together (there goes my lunch). An actual sciency post like the recent one on the history of violence can take several hours, what with finding and reading sources, writing, making illustrations and so on.

Some people claim an effective blog needs to post one big piece and a couple of fillers every day. That's too much for a science blog, but even at my current rate of three posts a week I'd need an extra hour every working day. That hour comes from my work time or it will have to come from my Japanese language studies, from reading, from photography, from spending time with Ritsuko. In short, from enjoying my time away from work.

If you want scientists in general to spend time on communication then you have to pay for it — pay by adding it into the job description; pay by setting aside time for doing so; pay by offering support and guidelines; and pay by offering training for those who aren't naturally good at or interested in social media or non-specialist writing. Understand that it would be quite doable; you could make part of your yearly teaching load for instance. It is teaching of a sort, after all, not just in a lecture format and not to a small, captive audience. It would mean somewhat less other teaching, or less research. That is unavoidable.


But do we really want to force all scientists to communicate directly? We don't require every public servant to communicate with the public in other fields, do we? It's rare for actual policemen, physicians, soldiers or firemen to communicate directly with the public, for instance. National administrators and bureaucrats don't tweet about new regulations all by themselves. They go through their public relations department and a dedicated press spokesperson.

Communication is hard — hard enough that public communication, journalism and outreach is an acknowledged group of professions. Why do we have this idea that every single scientist needs to do this by themselves? And, may I add, without the training, interest or time you'd need to make a decent job of it. Having stressed PI's dutifully dump their unedited abstracts onto Google+ or Facebook, or tweet the paywalled link to their latest paper, then ignore any questions and feedback would hurt science communication, not help it.

Reference Management Software

"Reference Management Software." Just listen to that noun; it's dull and dry as a bone. But, for those of you who don't know it's a hugely important tool for researchers. "Reference management" is all about being able to find research papers, to organize and sort them, and to create references for your own papers. Here's an interesting comparison of four RM systems.

For those that don't know, though, what are they and how do you use them?

Imagine, if you will, a research paper. This one, for instance, about the distribution of city sizes¹. Now imagine you're working on a project on cities and stumble onto this one. It may beome in useful so you save it for later. You drop it into your reference manager — RM — software; it saves the PDF file, and records the name of the paper, the authors, the date and place of publication, when you added it, and perhaps also keywords, what webpage you got it from and other relevant information.

It's some time later and your project is picking up steam. You vaguely remember some paper about city size you saved earlier; or you're looking for all papers by Ethan Decker; or looking for any paper that mentions Zipfs law — all of which would match this paper — so you turn to your RM and use its search function to find the paper again. You could search for "Decker", or "Zipf", or "PLoS city" and get all your saved papers that match. You find the paper and look through it, and as you read you add tags and notes about the paper right in the RM itself.

It's later still and we've read a lot, done our own research and we're well on our way writing our own paper. We need to add citations to all our sources in the text, and make a bibliography at the end. How you do that varies depending on how you write. If you use Word or OpenOffice you probably use a plugin that helps you search and insert citations from your RM. If you use LaTeX (and if you write research papers you should. Really.) you can simply export all your citations into a file that LaTeX will be able to read. In either case you're saved from having to type in and format all the citations and reference list manually — if you've never done that by hand I can tell you it's long, dreary work that is hard to get right and a pain to keep up to date.

 

OK, but why can't you just keep the papers in a folder and just get them from there? You could, if you have just a few dozen papers. You could be more methodical and save each paper by author and title, then store in a separate folder for each journal and year. But eventually it becomes unmanageable. I have several hundred papers saved just for my current project, and there's people out there with tens of thousands of papers stored. There is no way you can remember all the relevant papers you have, never mind actually remembering any particular details about each paper.

A reference manager helps you not only store your papers. A good manager helps you get them and add all the metadata it can without your assistance. It helps you search for and find relevant papers when you need them, and lets you add comments, notes and tags to them so you can find them again and remember what they were about without having to reread them. And it helps you generate and format your citations and reference list in a consistent, correct fashion.

For what it's worth I use Zotero as my manager. It's cross-platform — it works anywhere Firefox does — and free and open to use. The review above does a pretty good job of showing its strengths and weaknesses.

What I like most about it is the ease of importing and exporting data. When you find a paper online you just click on the import icon in the browser bar and the paper and citation data is automagically downloaded and indexed. It doesn't always work; for a few sites it only collects the citation data and you have to add the PDF file yourself, but it's not a major hassle. Exporting is also really easy. I just add papers to a collection specific collection for the paper I'm writing, then export the entire collection as a BIBTex file that LaTeX can read and use. It really can't be simpler than that.

If you're a budding researcher and doesn't yet use a reference manager you really owe it to yourself to start today. Pick any one you like, of course, but I think Zotero is a good choice if you don't know which to pick.

--

#1 It's a neat paper, and I've been meaning to write something about it. Very shortly, they show that while larger city sizes are described well by a power law, regional and smaller communities are better described by a lognormal distribution. They show how a very simple model of human migration and reproduction can generate the real distribution of smaller communities.

It could mean that network effects — that we want to live in a city because other people do so already — only kicks in over a certain city size. Below that size the number and distribution of communities depends mostly on local random migration and reproduction patterns. Smaller communities do not inherently attract people the way larger cities do. Is the paper correct? I don't know, but it's very interesting.

Linux Runs Computational Neuroscience

What OS is actually used in computational neuroscience? A recent paper in Frontiers in Neuroinformatics (let's hear it for Open Access!) has looked at this.

There's plenty of data there, but the main finding is that the most used system is Linux. Most researchers in the field use more than one OS, but Linux is the most common system, used by more than two-thirds of respondents, with Windows in second place with half and OSX third with a quarter¹. Some people use Linux as their primary OS while others use it in a virtual machine or logged in to a remote machine somewhere else. Of course, many, even most people use more than one system.

One reason for the popularity of Linux is that many computational research tools are developed primarily for Linux and Unix; another one is that clusters and supercomputers mostly run Linux today. If you need to run your model or computation on a larger cluster you will need to use Linux in one form or another. But it's not simply a matter of necessity; the paper finds that satisfaction is also highest for Linux. Windows is most likely to be used specifically to access Word, Outlook or other specific software that only runs on Windows.

Now, the paper is based on an online survey and a self-selected sample of respondents; this is problematic at best. But it does fit with my own anecdotal experience. At OCNC I saw some people that primarily used Linux, but many more dual-booted Linux and another OS, or combined more than one OS using virtual machines. I often see similar setups at conferences and meetings as well.

Virtual machines have long been used on big servers and mainframes, but are fairly recent in the desktop world. A virtual machine — a VM — is a piece of software that emulates a real computer. You can install an operating system and applications in it and the OS will think it has direct access to the real hardware. In reality the virtual machine runs as an application in a host operating system, and tightly controls the access to the real system.

A VM is extremely convenient. You can start and stop the system it hosts at any time; you can save the entire system state into a (large) file on disk and go back to that saved image whenever you want, or copy that image to other computers to run there. It lets you create a specific software environment guaranteed to be the same every time you use it. Modern PCs lets a VM give out controlled access to the real hardware so there is not much speed reduction.

You can use a hosted Ubuntu Linux system for your software development and data analysis, use a remote cluster for your actual simulations, and the desktop OS you're already familiar with for email and web surfing. Or run Linux or OSX as your primary system, then a copy of Windows in a VM to access legacy Windows-only applications. Or run a second copy of your system in a VM, to make sure the environment is identical every time you run a simulation.

 
The major drawback of the virtual machine approach is really that each hosted OS really needs as much memory and disk space as if it was the only system on the computer. But modern laptops tend to have plenty of both, and for large simulations you're likely to use a remote cluster anyhow.

--

#1 This seemed a bit low to me at first. But this survey counts desktops and clusters as well, not just laptops, and OSX isn't nearly as prevalent in those areas as in portable computing. Also, Apple laptops have a very distinctive, uniform design; you end up with a positive bias where you remember seeing them but forget about all the anonymous, generic laptops that were really the majority at the meeting or the conference.

Reality Isn't Anybody's Bitch— Threats and the Need for Pseudonyms

Why would you want to have a pseudonymous identity online? How about this reason: you and your family get threats from chronic fatigue syndrome sufferers because you show that it isn't caused by a virus. Same thing happens to researchers that show that the link between autism and vaccines is false¹. Same thing happens to climate researchers that show warming is, in fact, happening and is, in fact, caused by human activity. Geologists, palaeontologists and archaeologists that publish research contradicting some religion or another. Physicians that provide abortions, contraceptives or even just medical reproductive advice to women.

People that get targeted in this way can't have a normal online life. If you ask them to use their real names online, they will get tracked, harassed and shouted down whatever they try to do. Their families, friends and anybody in public contact with them put themselves at risk for the same kind of harassment. And while most threats are idle, a few are serious enough; researchers and their family members do get injured and killed in attacks.

But even idle threats and harassment is a serious thing to those who get targeted. More and more, our public discourse is online. If you get shouted down, if you get chased out of that discourse, your right to participate — and our democratic systems are built on people having that right and using it — is compromised, and your viewpoint goes unheard. Googles shortsighted "real name" policy is ultimately very damaging, to Google itself, but also in a small way to the greater society.

 
Real names don't stop this harassment. Most people harassing scientists already do so openly, under their own name, proudly brandishing their membership in whatever cult or organization tells them to go ahead. Real names forces away the victims, not the attackers. Real names don't stop mob rule; it legitimises it.


But why are scientists targeted? This is not some temporary phase, and it's nothing new. It is because reality always wins.

You can't wish away reality. You can't ultimately ignore it. You can't bribe it off, reason with it or make a deal with it. You can curse it; condemn it as immoral and evil; you can make it illegal — but reality just Will. Not. Care. If your ideology or religion contradicts reality, well so much the worse for you and your ideas. Reality won't shift just to accommodate you.

We all live with a complicated structure of ideas, assumptions, prejudices, ideologies and notions of how the world around us ought to be. We are not likely to like people that come along and show that bits of that structure is wrong or even harmful.

The Chronic Fatigue sufferers wanted a cause — any cause — for their condition, even if it doesn't lead to a cure. A few no doubt want it to be a physical cause; mental conditions are still heavily stigmatized and it becomes so much easier to bear when you can point to something physical. A chronic virus infection is a great explanation. It fits a lot of the data and you finally get some kind of target; it points to things you can try and ways to alleviate symptoms, but most of all it gives you a coherent reason. We humans love to have reasons for things.

Then a bunch of researchers comes along and shows the data was wrong — there is no virus and we still have no clue. Some of the patients feel it really would have been better to leave things alone; a comforting lie is easier to live with than a disagreeable truth. Getting hope, then having it yanked away from you breeds a lot of resentment, denial and even, in a few already unstable individuals, threats of violence. And since reality is inviolate that anger, resentment and violence gets directed at the messenger instead.

People realize their cherished ideas are safe only as long as nobody shows they contradict reality. So the only way to keep their ideas safe is to drive out, shout down and silence those that would expose facts that contradict them.

This, I suspect, will become more and more common, and spread to more fields over time. There is hardly a field of research that doesn't contradict some dearly held beliefs of people somewhere. Through the internet those believers can now easily find each other — and find researchers that publish contradictory findings. It's much easier to disrupt somebody's online presence than their real life. At the same time, your online life is rapidly moving from an idle hobby to an important part of your core life.

I write very little about my work online as it is. If I were in any kind of sensitive field, I doubt I would write anything science-related under my own name at all.

--

#1 And in that case, the initial claim was deliberate paid-for fraud on the part of Andrew Wakefield, committed in order to bolster legal malpractice cases and to give a market opening to his own, alternative vaccine.

Is Internet Destroying Your Brain!?- hint: No, it's not.

Rewire Your Brain
For Fun And Profit!

Susan Greenfield, a supposedly actual scientist — somebody who really, really should know better — ignores any contrary evidence in order to push a scare about the internet destroying young people. She's the source of the idea that "The Internet is Rewiring our brains!!! (OMG! We're going to DIE!!!)".

Well, of course using the internet rewires our brains. Any experience rewires our brains — that's what brains do. When you remember something, that happens by the brain rewiring itself to add that memory. You can use the same argument for any experience: "Having sex rewires our brains!", "Having breakfast rewires our brains!", "Having fun rewires our brains!". Reading about Greeenfield certainly rewires my brain to become cranky and annoyed for hours afterwards.

She's been pretty thoroughly debunked, but lack of evidence doesn't seem to stop her in any way. She's apparently not happy with the waning press coverage however, so now she's claiming that the internet and social websites in particular is a cause of autism.

 
Several people such as Dorothy Bishop point out some obvious problems with that idea, such that a) the rise in autism diagnosis started long before the internet became a public medium; and b) autism is typically diagnosed at the age of two, long before children start using social sites. Her answer — her full scientific argument, really — is:

"I point to the increase in autism and I point to internet use. That's all."

Apparatus Of
EVIL!

"That's all", and that's the problem. Carl Zimmer (read his blog; it's great) Have an excellent summary. And as he points out, some people have taken this brilliant piece of argumentative excellence and run with it on twitter. The result is pearls like:

"I point to Alzheimer's and I point to cheese doodles. That's all."

"I point to diabetes and I point to cats. That's all."

Take-away lesson: Somebody may have a "PhD" at the end of their name — or an "MD", or a "Rev.", or "Civ. Eng.", or anything — but that doesn't guarantee they're not ignorant about what they say, or that they're not trying to push an agenda they know to be false (if they're a "Rev." that's more likely than not of course).


Sometimes an expert in one field wades in to do cocksure pronouncements in another without understanding it, and end up embarrassed. "Expert" critics of global warming, for instance, tend to be engineers or physicists without a background in climatology. Engineers and physicians seem particularly prone to this. They have a method-based hands-on understanding of one particular subfield, and think they can apply their methods to other fields without learning enough background to actually understand it.

 
And we're all human. As much as we want scientists to be objective and disinterested, the reality is that we all have our biases and hobby-horses, and we're concerned about our own careers, our pensions and mortgages. We grow old and cranky and stop learning new things. We refuse to let go of favourite ideas long after they've been disproved. We may have religious or political biases that blinds us to some facts. Some may see fame and fortune — or at least a modest side income — from ignoring evidence and science in favour of publicity and a book deal.

I'm not arguing against expertise. An expert is a lot more knowledgeable than a non-expert and people know this — would you rather have a bypass operation done by a heart surgeon or by a plumber? But check that the expert really is an expert in that particular field; how about open heart surgery by a psychiatrist, a dermatologist or an eye surgeon? They're all MDs too after all.

Ask yourself if this person might have a bias or some reason to not be completely open and accurate. If they happen to run for office, or if they are devoutly religious they may have reason to make the science sound like what they or other people want to hear. If they're selling a book, or if they're paid by a particular interest group or company there's reason to be suspicious.

But most of all, if a claim is at all sensationalist or surprising, ask for the evidence. If they have evidence — if they have data and analysis, and if that has been corroborated by others — they should be happy to disclose it. If they show only sketchy, suspicious or discredited data, or if they show nothing at all, then that probably means they have no solid evidence. If so, ignore them. If they are right, then people will find good, solid evidence sooner or later. If not, then you won't have wasted your time, and possibly your health and your money, on a fraud.

I Submit

I've been working on a paper off and on for most of this year, and last week we finally managed to submit it. Now we wait — we wait for the editors to decide if it's appropriate for the journal; we wait for them to send it out to reviewers; we wait for the reviewers to submit their reviews, and we wait for the editors to decide whether the paper has a chance of being accepted.

This can take anything from a month (that would be fast) to six months, though a year is not unheard of. I'd expect it to be three or four months. If they decide it can be accepted, we have to revise the manuscript according to the reviewers comments, or explain clearly why we think a suggested change is unnecessary or harmful. The revisions can range from spelling errors, up to redoing the whole model from the ground up, running new sets of simulations or completely change the way we analyse our data. Our changes may be submitted back to the reviewer for further comment if the editor thinks it's needed.

If the editor or reviewers think the paper is not acceptable, or if we don't think the required revisions fit the paper we want to publish, then we give up on that journal. We'll decide on a different journal, rewrite the paper to fit that journal, and submit it again. All in all, six months to a year from first submission to publication would be quite normal. Three months is fast; expect a year and a half to two years if you have to resubmit the paper.

 
All well and good — except that our project ends next March. We lose access to the cluster computer we've been using for our simulations, and I no longer have a job. We might come to the point where we're asked to do a new set of simulations for the paper and we simply can't: we no longer have the computing power, and I might not even have a science-related job any more so I may have little or no time to work on the model or the paper.

It won't come to that, hopefully. Even if the project ends, we could probably ask for a little computing time to finish the project. And if I find a research-related job it's accepted practice to spend some time finishing up things from your previous projects. Time will tell, as always.

Hauser Resigns After Research Fraud

Marc Hauser has resigned from Harvard, following the discovery that he falsified experimental data in at least three papers. He is going to work "on the educational needs of at-risk teenagers" (which is really laying it on with a trowel; couldn't he just "want to spend more time with his family" or something?)

Why would a lauded, even lionized researcher of such stature commit fraud? Only he knows, of course, but I suspect his towering status has more than a little to do with it. Author of multiple books, leader of his own lab in one of the most prestigious research places in the world, greeted as a celebrity wherever he goes, half his own staff and students seeing him as a hero - the pressure to perform, to keep getting results, must have been immense. No matter what, he is still only a fallible human, and asking him to be superhuman, to be perfect, is not realistic and not fair.

Very few people are born cheaters. The vast majority of researchers enter their field because they genuinely want to understand the world around them, and share their finds with the world. If money and fame is your goal then science is not the right field for you. I doubt most cheaters start by outright, large-scale fraud, and I doubt Hauser did either.

But I imagine that when failure is no longer an option - when your career, your status, your own livelihood and the future careers of people under your charge all depend on your success - people will all too often start down a slippery slope: clean up a fuzzy picture just a little, omit those obvious outliers, perhaps rerun an experiment that just failed to reach significance. The pressure doesn't let up, and once you've started it's easy to take just another small step, and another... One day you wake up and realize you've become something you used to despise.

How to stop this? We can't stop hero-worshipping, and research funding agencies understandably want to support successful research over failures. It's one thing to tell people that it's OK to fail, but the reality is that a failed project is a real handicap when funding is as cut-throat as it is today. Better oversight and a culture of transparency would help - it should always be OK to talk about odd events or suspicious data in your lab, and a whistleblower should ideally be able to count of the full support of their university. Reminding people what is and is not good research practice throughout their training is another good idea; almost all people want to do things correctly, so you want to stop them from entering that slippery slope before they even realize it.

 
As for Hauser himself? He may or may not return to academia in some years. But with this resignation - and the lingering doubt of all his earlier publications; he is unlikely to have started with full-blown fraud - I don't expect him to ever return to active, high-profile research ever again.

Don't Wait Up for the Singularity

Ever hear about "the singularity"? It's the vague idea that technological advancement and machine intelligence is exponential, and will completely overwhelm humanity. One part of that is the sci-fi idea that we'll be able to analyse and copy the exact brain pattern of an individual, then have them live on as potentially immortal simulations, perhaps with robotic bodies to interact with the real world. Standard science fiction fare, but some people believe this is a real possibility.

In this piece, David Linden argues that while it is conceptually possible to actually copy somebody's mind, the proponents are very wrong on just how difficult it is, and how much time it will take before we'd be able to actually do this.

As a computational neuroscientist (well, sort of at least), I agree with pretty much every word he says. We are a product of our brains, situated in the world, and if you were able to do a faithful functional copy of the mind and of the world you'd be able to live on. No ghost in the machine is necessary. However, we are very, very far from being able to do anything like this. Never mind knowing enough about the brain to do a faithful functional copy; we don't even know enough of its function to understand what we'd need to know in order to do this.

 
There are people bandying about a time frame of 10-20 years for this kind of thing; a more realistic estimate would be anything between ten and fifty times that number. I certainly don't expect to see a successful experimental result within my lifetime.

Too Many Papers

Time for some bellyaching. I have a collection of RSS feeds from various research journals that show me recently published papers I might be interested in. This morning I had 208 new papers waiting for me.

208 papers that might interest me. In one day. Just from the small subset of journals I have RSS feeds for. I spent an hour just eyeing through the titles, and of those 208 papers there were five that I clicked through and will take a real look at.

First, this tells me I need to set up better filtering. Most papers are of no interest to me at all, but most journals will simply push a list of all their published papers and not give me a choice of what to see. The problem with filtering is of course that I might miss papers that turn out to be relevant. I can live with that.

Second, it is incredibly annoying when a journal RSS feed contains only the title, not the abstract - "Nature", "Science", "Nature Neuroscience", "Journal of Cognitive Neuroscience" and "Neural Computation" are all guilty of this. If I have 200+ new items to go through then I am not going to click through every promising title to find out what the damn paper is actually about. I end up only clicking on the papers that are obviously relevant and ignoring the rest. Yes, I may miss relevant papers, but I can live with that.

Oh, and a special mention to "Science", which has no feed with only the new papers. Instead they see fit to publish only an RSS feed with everything in their latest issue. I'm not very inclined to read editorials, "Science 50 years ago" or blurbs about new books or exciting new laboratory equipment when all I want is to dig out from under my morning pile of papers. And sometimes they publish their RSS feed before the corresponding web page is actually available. If I get an error page when looking for a paper then I am not likely to remember to try again tomorrow. In fact, the "Science" feed is annoying enough that I consider removing it. I may miss relevant papers that way, but I can live with that.

 

How can I live with missing potentially relevant papers? Because I'm drowning in relevant papers already. The challenge isn't lack of information; the challenge is to not get overwhelmed by all the new information washing over me every day. As I said, I found five potentially useful papers this morning; that's a typical number, and I might stumble on to another few before the day is over, without even looking. Just a quick look at the new papers, perhaps follow a reference or two of theirs in turn and I'll spend most of my morning catching up with new papers. A focused search for information will easily eat up the rest of my day.

I'm very unlikely to miss something truly important. I'll hear about any ground-breaking new paper soon enough. But the vast majority of papers aren't ground-breaking; most are barely of interest to anybody but their authors¹, and is it's relevant to me I'll likely stumble on to the same info in some other paper later on.

This flood of papers keeps getting worse. I - and everybody else - will have to continue to narrow down our interests, to drop sources, to become quicker and more ruthless with out sifting, and - yes - missing things in the process. If we don't, catching up with current research will leave us with no time for any actual research of our own.

What will go? If my own experience is any guide, any journal that makes finding and reading their papers difficult will lose out. Anything I can't find through normal searches; any publication with restricted online access; any paper I can't download directly. If it's truly ground-breaking, then I'll hear about it. If not, I'll just find the same information somewhere else.

 
--

#1 Yes, this goes for my own work as well. Most of science is about crossing t's and dotting i's, and makes very little direct difference for anybody else. That doesn't mean it's useless; it's all work that needs to be done, and the rare, high-profile breakthroughs depend on it for them to happen. This should perhaps really be the subject of a separate post.

DIfferential Gears

I was going to write something else before the weekend, but I just stumbled on this absolutely genius instructional video on how differential gears work. I've known the point of a differential - give the same force to two wheels, even though they rotate at different speeds - but I've never quite wrapped my head around exactly why it works. This video made the trick.

I love almost everything about it - the black and white tones and the lighting is excellent; the examples and explanations are exemplary, though exceedingly eloquent. The one slightly annoying thing is the narrators cadence, which is just a little over. the. top. as. he. empha. sizes. every. single. word.

The video doesn't bring it up - it's about automotive engineering - and I'm not going to belabour it, but if you control the center piece instead of letting it rotate freely, you can use it to set the relative speed and the phase of rotation between the two output axes. Me and a colleague had a vague idea for using that to create a mechanically simple four-legged walker, but as so often we never found the time and motivation to actually do anything with it.

Nukakas Wedding

I'm still working on The Paper That Will Not Die, and it's draining my enthusiasm for any other kind of writing. Actually, it's draining my enthusiasm for anything, period, and I'm questioning my choice of career¹ at the moment. I have daydreams of working as a ramen cook or convenience store clerk; the pay sucks, but the hours are shorter and they don't need to write a paper about their work every few months.

Anyway, I found what seems to be an amazing book at another blog: Nukaka no Kekkon - Nukakas Wedding. The author takes insect behavior and translates it into human fairytales. The resulting stories, written in a children's storybook style, have an almost otherworldly feeling, and they're illustrated in a simple, clean style that goes really well with the stories.

One story is online here: The 100 suitors. It's in Japanese, but only ten pages in total, each one illustrated and with just a short text per page, followed by a page explaining the original insect behavior. Translation below:

1/10 One day, a beautiful woman appeared on a hill overlooking the village.

2/10 It caused an uproar among the men in the village. "Please marry me." "I'm the man for you." "Please choose who you'll marry!" A hundred men were asking to marry her.

3/10 "Please don't worry" she answered, "I'll marry all of you." "Follow me, those of you who want to marry."

4/10 The men jostled to follow her as she left.

5/10 As night fell, they came to a cave high in the mountains. "Here's our wedding place. Please come in."

6/10 As they entered the cave, the men asked impatiently "How on earth are you going to marry all of us?"

7/10 "Like this" she said, and breathed a white mist all over them.

8/10 The men all suddenly turned to stone.

9/10 The woman would turn one man back from stone each year, and marry him for that time.

10/10 Explanation: Many ant species will mate only once in their lifetime, but they will do so with many partners. Winged young queens will mate in flight with as many winged males as possible. The sperm isn't used at once though; instead it is stored in a spermatotheca in her body where it can hibernate for years.

The queen flies off to find a place for a nest. The hibernating sperm is activated, one bit at a time, as needed to fertilize her eggs. In this way ants can mate only once, yet carry the offspring of hundreds of partners. For example, the Japanese wood ant can carry 5000 eggs throughout her 15-year lifetime.

If this free story is any indication it seems like an excellent book. It's out of print, unfortunately, and used copies go for more than I'm really willing to pay. If it gets a reprint, or if I find a used copy at a reasonable price, I'll definitely get it.

Insects are a different world unto themselves. Kind of wish I'd studied entomology at university when I had the chance.

#1 Well.. what could pass for an actual career in fog, at night, at a distance, if you squint, don't look too closely and get distracted by a circus elephant riding a unicycle at the right moment.

JoUR

So, your research project is on a tight deadline. You're scrambling to put together a paper and push it out to some journal in time for a looming project review. You're hunting references, writing and rewriting your arguments and doing a desperate last-ditch data analysis round in the hope this will be enough to find favour with the journal reviewers. Meanwhile the clock is ticking... tick. tock.

Let's face reality here - you're out of time and out of luck. Your paper looks like what it is, an unfinished cut-and-paste job from your earlier publications, with some obvious editing to make it look vaguely coherent. This really has no chance to be accepted; you'll get a "revise and resubmit" if you're lucky, and an outright "reject" if not. Either way, you'll be spending a month or so rewriting it from scratch.

So why not go a little easy on yourself? Stop panicking and stop wasting precious time you can't afford. Just take whatever you have and submit the unfinished, unreadable mess to Journal of Universal Rejection.

Skip the last-minute editing, the tedious formatting, the Byzantine submission guidelines and the wait in the endless submission queue of other journals, and go directly to the inevitable rejection. The same end result, but weeks - maybe months - faster! Think of all the time you'll save! You'll have your revised, rewritten manuscript ready for submission to a good journal before your colleagues have even received their rejection notifications!

 
Remember: "Journal of Universal Rejection - when 'submitted' is what matters."

The Joy of Stats

Here's a treat for you: An hour-long television program on statistics.

Wait! Don't run away! First, it's a BBC production: excellent production values, good pacing, great footage. Also, the host, Hans Rosling, is exceedingly energetic, engaging and entertaining as he euphorically enthuses about how statistics can describe our world. This is easily worth one hour of your time. I've never liked statistics as a student, and I watched this straight through.

For a quick taste to see what you're in for, take a look at Rosling when he describes 200 years of the world in four minutes.

Update: changed the first link; now it hopefully works for everyone.