2020: A Moderately Bemused Review of the Year in Neuroscience
That Was The Year That Was
It seems we are cursed to live in interesting times. A year starting with the threat of conflict, Donald Trump authorising the assassination of the Iranian General Qasem Soleimani, Iran retaliating with rocket attacks on US military bases, and then tragically shooting down an airplane of its own citizens by mistake. All banished to distant memory by the global pandemic unfolding in slow motion.
And yet ironically I’d wager for most of us this year has been the least interesting on record, a year in two dimensions, endlessly staring at a screen in your own home, reading from it, typing to it, sometimes watching it, sometimes observing the tiny heads of people you once knew as fully three-dimensional beings saying “sorry, you go first” back and forth until you can barely resist the urge to hurl the screen through the window and run for the hills.
Somewhere in all this, neuroscience still happened. So read on for a look back, a look forward, and a little look at my new book.
The world moved online, and neuroscience was no different. The Neuromatch conferences led the way, kicking off in early May with a computational neuroscience meeting, and graduating to a full-on replica of the (cancelled) SfN Annual Meeting in late October, right down to the titles of its “Themes”. WorldWideNeuro set up a one-stop shop for all the one-off seminars and talks happening online around the world. Much debate about whether these were better or worse than doing stuff in person. They are both, of course: better access for a more diverse audience; worse for actually meeting people and doing the kinds of things that create bonds between people. But there’s little doubt that how we disseminate our science will be changed forever, with online, virtual events proving their worth.
The Neuromatch Academy, an online summer school for computational neuroscience, was truly an awesome achievement. The sheer scale and quality of the content was breathtaking — and still is: you can access it all, for free. As was the sheer scale of the number of faculty involved. My involvement was peripheral, a mentor for two groups working fiercely hard on their projects, but was still a lovely little vignette of how the Academy opened up the world, me in the UK mentoring one group in China and one split between Australia and New Zealand. Would be wonderful to see it again next year, if any of the organisers remain capable of seeing the words “Neuromatch Academy” without their eye involuntarily twitching.
Fascinating neuroscience research kept appearing print, oblivious to the world outside. Special mention for the paper from Simon Peron and colleagues in Karel Svoboda’s on the local circuits in cortex, for two remarkable things. For one, being a proper, straight-up test of causality in systems neuroscience by identifying single neurons with strong coding of the outside world, and then specifically killing each of them to see if they really were carrying the code for the population. And for another, the computational model driving the research is Figure 1 of the paper! This was one of the few neuroscience papers I’ve ever read that looked like hard science as we aspire to do it: derive predictions from a model, and causally test them in experiments. Special mention too for Srdjan Ostojic, his lab, and their collaborators (including Omri Barak) who seemingly overnight unleashed on us an entire research programme on when, how, and why networks of neurons will have low-dimensional structure, giving compelling answers to some of the deep questions of systems neuroscience.
Some chilling papers appeared too. The NARPS paper showed us how seventy teams of analysts looking at the same set of fMRI data agreed on almost nothing about which hypotheses those data supported. As I wrote earlier this year, nothing about the failures in this paper were specific to fMRI research — the authors were showing us how any area of science with complex data processing and analysis pipelines could have teams of brilliant, smart people end up at polar opposite conclusions from the exact same data.
And a stuff-of-nightmares paper for any modeller showed that a widely-used model for estimating the dynamics of the evolution of existing species could produce an infinite number of evolutionary trees consistent with collected data. So any inferences made from that model are garbage. Cue a bonfire of papers, and computational people from all other disciplines muttering “there but for the grace of god go I” under their breath.
COVID-19 research didn’t have sole claim to a tranche of wonderfully terrible and over-hyped papers this year, of course — the lure of scientific fame and glory still brings out the overreachers and the just plain bonkers, and neuroscience was no exception. We learnt that crows have consciousness apparently, just because when they report seeing a faint grey square we can detect that imminent report in their neural activity. Which amounts to “brain has memory of seeing a thing”. Which we already knew ought to be in their brains, because if that activity wasn’t somewhere in their brains then they wouldn’t make that decision. Unless crows use their spleens to make choices. Now that would be worth publishing in Science.
And fittingly for a year spent for many of us in a state of repose, it was a good year for reading about neuroscience. Matthew Cobb’s “The Idea of the Brain” kicked the year off with its deep scholarship of the history of neuroscience worn lightly and entertainingly, and ended the year as The Times (of London) book of the year in Ideas & Philosophy. Serious brain books doing seriously well. Ashley Juavinett’s “So You Want To Be A Neuroscientist” debuted in November, and its lucid, lovely guidance should be immediately gifted to anyone you know who is already a neuroscientist, desperately wants to be one, or once remarked in passing that brains are kinda cool. And your parents, so they can find out what it is you do for a living.
The outline of 2021 is vague. Vaccines have arrived, a superhuman effort of teams of scientists and regulators and manufacturers around the globe. Their mere existence does not promise a return to pre-pandemic life: for one thing, a global pandemic requires a global vaccination programme, and right now the richest countries in the world have bought every vaccine they can lay their hands on. But by end of 2021 it seems each of our worlds will have the potential to become more three-dimensional again, to teach and meet freely in person.
We could, but will we? It seems unlikely we will return to an exact replica of pre-pandemic life. Much of science and of the life of universities will stay online, for having been forced to finally make that painful step under duress, we now have new ways of disseminating and teaching that can be more diverse, more inclusive, more free.
No doubt we’ll see something mind-boggling from DeepMind, after AlphaFold near-as-dammit cracked the prediction of how proteins fold into their three-dimensional shape. My money is on AlphaOmega, a Deep Learning natural language network, trained on a combined data-set of world-wide media and every work of science fiction, that will accurately predict exactly which combination of idiocy will end the human race.
We’ll have a better idea of the future shape of how science will be published. The preprint revolution gathered extraordinary pace in 2020, as preprints on all aspects of COVID-19 research became the main currency of communication, the lifeblood of media coverage and government policy alike. The journal eLife is about to make preprints mandatory for all submissions, turning them essentially into an “approver” of work already published. Physicists meanwhile continue shaking their heads at the rest of us, wondering what took the rest of science so long. But we can at least take comfort in the fact that we’ve caught up quickly.
Talking of publishing… Psst, keep it to yourself, but I’ve got a one-time, never-to-be-repeated deal to offer you for 2021: give me several thousand dollars and I will write an illogical, error-strewn rant about how crap your life’s work is, in prose that was either written at midnight under the influence of a triple espresso or by a narcoleptic mountain goat.
Tempted? That’s literally what the Nature family of journals have announced as this year drew to a close: should you aspire to publish an open-access paper in Nature Physics, Nature Methods or Nature Genetics, you pay them 2190 euros ($2650) upfront for the privilege of reviewing your paper. Just reviewing. Not publishing — oh no, that costs lots more money. Just reviewing. And an editorial suggestion about which journal in the Nature family your paper is best suited for, if any. So not only do you now get to pay for this abuse from the reviewers of your paper, you pay to be rejected too. What a bargain.
This outlandish proposal is a ramification of Plan S, that bold proposal from the European Union to ensure that all papers from grant-funded research are immediately available to read by anyone no matter where they are published. Journals whose life-blood are the subscriptions paid to them by universities around the world are expected to find new ways of working, of generating the cash needed to keep them afloat. Nature have already given us a glimpse of the kind of nightmarish logic such journals will now be caught up in. As 2021 unfolds, we’ll see more such craziness, and in reaction more and harder pushes for alternatives to the long con that is academic publishing. [For more on this, read my longer take].
Neuroscience will, at least, be entertaining in 2021. This year saw the launch of In Silico, the documentary movie of the slow-motion disaster that is the Human Brain Project. Launched at a New York film festival, and well-reviewed in Nature, with luck 2021 will see it available to rest of the world.
Grace Lindsay’s book “Models of the Mind” launches early 2021, promising us a manifesto for how theory has illuminated every corner of neuroscience, from how the brain makes decisions, to how its neurons are wired together. Can’t wait.
And, of course, the thing I’m most excited about for 2021 is the March 2nd launch of my book “The Spike: An Epic Journey Through the Brain in 2.1 Seconds”, a rip-snorter of a journey through your brain. Its goal is simple: to understand ourselves, we need to understand the language of our brains. And our brains talk in spikes, blips of electricity passed between our 87 billion neurons. The book is a spike’s journey through the brain, from eye to hand, revealing their mysteries, their meaning, and a new vision for how brains work. I hope it’s as much fun to read as it was to write.
Good luck all, and see you in 2021.
Want more? Follow us at The Spike
Twitter: @markdhumphries
