I am a 4th year PhD student, awaiting my viva. My thesis was an electrophysiological investigation of the sources of input to dopamine releasing neurons to help understand their function. In the meantime, I've worked for an environmental charity, creating teaching materials for 5-11 year old children to accompany a project to promote urban and suburban bee keeping. More
Inspired by the statistics classes I’m currently teaching, and a conversation I recently had in the pub with some colleagues (because I’m just that exciting), I’ve been wondering about why p < 0.05 is the most common threshold for statistical significance, at least in the psychological sciences. I realised that the choice of threshold was probably arbitrary to a certain extent, but I thought that maybe it was at least a useful arbitrary value for whatever purpose p values were first used for. I had been teaching about t-tests, so they were on my mind. I knew that the Student’s t-test was created by William Gosset to help quality control at the Guinness brewery (the brewery forced him to publish under a pseudonym – Student – to conceal from competitors that they were using statistics). Perhaps a false positive rate 1 in 20 was considered to be a reasonable error rate in brewery quality control? Apparently not…
The threshold, or indeed any threshold, doesn’t seem to have arisen with Gosset. P values certainly pre-date Gosset and the t-test anyway, but the publication of his tables of the t-statistic (or rather, what he referred to as the z-statistic), and the tables of his colleague Pearson’s χ2 distribution, provided precise p values to 4 decimal places for a given value of t or χ2. Instead, our fixation on p < 0.05 seems to be at least in part due to the issues between Pearson, and another statistician, R.A. Fisher. Fisher had created more statistical tests, and wanted to reproduce Gosset’s tables. However, permission was refused because of financial issues over granting copyright, and disagreements over theory between Pearson and Fisher, and Fisher had to re-create the tables.
Sections of Student's (top) and Fisher's (bottom) tables. From Clauser (2008) Chance, 21(4):6-11
Fisher rearranged the data, and instead of providing exact p values for a value of t, he provided t values for values of p. It is apparently “a matter of historical fact that Fisher was the first to have published tables in this form”, although there is evidence pre-dating Fisher and Pearson that p values were considered as an indication of findings of further interest, usually around 0.05. Warnings about the overuse of thresholds of significance were also surfacing as early as 1919 – 6 years before Fisher’s tables, but the publication and widespread use of Fisher’s tables in a form that focussed on round p values probably helped reinforce the habit. Fisher doesn’t appear to have recommended the use of absolute thresholds of significance. Instead, he considered p values above 0.2 to be indicative of no effect, but values between 0.05 and 0.2 to be a suggestion that an effect might be detectable with sufficient modification of the experiment. Most of his tables reflected this; they provided values of several test statistics for a range of p values. However, for simplicity, his tables for the newly introduced F statistic were limited to p = 0.05. Although this was expanded in later versions, people seemed to have latched on to 0.05 as an important value.
Perhaps because the tables opened up the arcane world of statistics to a wider audience, or maybe because of some historical tendency towards 1 in 20 as an intuitive compromise of sensitivity and false-positives, Fisher’s tables seem to have left us with the one thing that everyone who knows anything about statistics ‘knows’. Maybe if Fisher and Pearson had been on better terms, undergraduate statistics might have been very different…
Clauser, B. (2009). War, enmity, and statistical tables CHANCE, 21 (4), 6-11 DOI: 10.1007/s00144-008-0004-8 Stigler, S. (2009). Fisher and the 5% level CHANCE, 21 (4), 12-12 DOI: 10.1007/s00144-008-0033-3
Also, see Gerald Dalall’s article Why P = 0.05? for more detail, or if you can’t access the papers.
Hello again! My viva is due on the 1st of December, in the meanwhile I’m looking for post-doc work. Something I’m thinking of going into is working with octopuses and/or other cephalopods – mainly their visual perception, their ability to camouflage, and perhaps how the the two systems work together. Although I can’t do them justice to the extent that my blog-mate Mike Lisieski at Cephalove can, I thought I’d write some of the stuff I’ve been reading about the fascinating squishy beasts.
Octopuses are part of the subclass of modern cephalopods (coleoids), along with their relatives the squids and cuttlefish. I had a browse through Binyamin Hochner’s “Quick guide – Octopuses”; the name suggested it was a good place to start. He says they diverged from old cephalopods (nautiloidea – which includes various extinct species, and the surviving Nautilus) around 450 million years ago, but that they “arrived in the sea at the same time as bony fish; more than 200 million years ago.” I’m not sure if that means they were fresh water before that, or whether the margin for error is just that big. Either way, octopuses have been around in one form or another since before land animals were land based. That’s a long time…
A common octopus (Octopus vulgaris) - from Wikipedia
Octopus intelligence
Given that amount of time, it perhaps isn’t surprising that octopuses seem to have evolved some form of intelligence. Intelligence is a difficult thing to define, but even comparing octopus abilities to a vertebrate benchmark is difficult given the radical differences between the behaviour of an inveterate aquatic organism and the behaviour of terrestrial vertebrates we are more used to examining. However, learning and memory tasks provide a basis for quantitatively assessing the cognitive abilities of octopuses and comparing them to other species. Octopuses have an amazingly wide range of behaviour, but notably they not only show operant and associative learning, but also long-term memory from visual and tactile tasks, and the ability to from observing other octopuses.
Octopus brains
What about octopus brains? The octopus central nervous system (CNS) contains about 500 million nerve cells, many more than the next most advanced vertebrates such as bees and cockroaches (~1 million) and not far off a dog (~600 million) or a cat (~1000 million). However, octopuses don’t have such big brains – about 3/5 of octopus CNS neurons are out in the arms (note: not tentacles – there’s a difference). These neurons are capable of producing complex, natural looking movement, even when disembodied. They are connected to the brain by a relatively small number of neurons, suggesting they send highly processed information and receive high level commands, rather than pushing huge amounts of raw data and receiving extensive direct control. This might well be the best way to control complex appendages that can bend at an infinite number of points, rather than the fixed joints of vertebrate limbs.
Masters of disguise
Cephalopods are masters of disguise, with an ability to change colour, and sometimes shape, to disguise themselves to a greater or lesser degree depending on the species. The degree of matching can be astoundingly convincing, and change in a fraction of a second.
Apologies to Dr Hanlon, who created the video, but I nabbed it from YouTube. The octopus goes from almost totally camouflaged to conspicuous in a few hundred milliseconds, then totally visible in a couple of seconds.
Cepahlopod camouflage is thanks to chromatophore organs – sacs of pigment with radial muscles attached around the edge. I had thought that chromatophores were not under central control, which makes their co-ordinated patterns even more incredible. They are, in fact, innervated from motor centres of the brain, although they are lower-order centres, and there is a one to one connection of motorneuron to chromotophore, so perhaps they aren’t much more co-ordinated than I thought. What is certainly surprising is that the colour imitation can be done by creatures that are colourblind (although they can be fooled by patterns designed to exploit their visual limitations). The texture effects of camouflage are thanks to changeable skin papillae, which seem to work on a similar principle to the arms (a muscular hydrostat).
More than just bright lights, big brains, and bendy arms
Obviously, there is more to octopuses and cephalopods in general than what I’ve mentioned above. They have some of the most complex eyes found in invertebrates, the ability to perceive and perhaps exhibit patterns in polarised light. The suckers are a whole world of fascinating features too – they change shape to provide an incredible amount of suction, and are packed full of tactile and chemoreceptors that give cephalopods a surprisingly detailed picture of their world. They are fascinating creatures, and I would love the opportunity to learn a little bit more about them.
There’s a rather nice infographic about the evolution of the eye by fellow SFSer Arthropoda.
The British Psychological Society have started Morsels – and addition to their research digest – covering recent research in little bite-sized chunks. It’s been running for almost a couple of months now. Go check it out!
Thinking of become a postdoc? Are you already a postdoc? Then the new Journal for Postdoctoral Affairs might have something for you. Written by postdocs, for postdocs.
Bradley Voytek at Oscillatory Thoughts collects together some of the top unanswered questions in neuroscience that some of the big names have posed in recent years. The post is a couple of months old, but I imagine some of the questions will still be unanswered! Head over there and join in the debate.
I’m back! The thesis is at the printers, and should be submitted on Friday. Viva is any time up to 10 weeks from then. I’ve read a few posts over the years about PhD experiences, but I think a) it’s still too much of a shock that it’s pretty much done b) I don’t have anything really to add, so I’ve no plans to do one at the moment.
What I will do is write about a blog post that I read in the tail end of my thesis writing that made me see things a little differently. It’s from xcorr, a computational neuroscience blog by the fantastic Patrick Mineault. A lot of it is well beyond my ken, but occasionally it’s provided some inspiration. This is one of those times.
Patrick had written up a research paper by Torralba in Visual Neuroscience that gave me some pause for thought. I won’t cover the ground again, as he did a perfectly good job already. The essence of the paper was that people can recognise indoor versus outdoor scenes from 16 x 16 images that had been blown up and blurred to produce a larger, but fuzzy image. To put 16 pixels into context, that’s the height of Mario in Super Mario Bros on the NES. That’s not a lot of detail.
Pictured: Not a lot of detail. From fantendo.wikia.com
Now, indoor and outdoor scenes could be distinguished based on colour; outdoor scenes had lots of lovely blues and greens, and indoor scenes were all those colours that weren’t colours until Dulux colour charts were invented. But it gets better: Participants could identify objects within these scenes. This was despite them covering perhaps a few dozen pixels each. Independent of their context, they would be unrecognisable, but within the context those few pixels could be identified as “probably a sink” because they make up something square, white, and in a bathroom. This leads to a bit of a circular situation where objects are identified based on their context, but the context is identified based on the objects it contains. It’s a bit vague, but I’d hazard a guess that the brain solves the problem by processing in parallel – simultaneously guessing context and object identity and updating as more information comes in.
The part of this that caught my attention was not how we identified objects, but that we could identify them based on very little detail, given a context. I’d be interested to see how long it took, and whether participants had to really study the pictures. Is identifying fuzzy objects a complicated process that needs interpretation by higher levels of the brain, or is it a heuristic, ‘best-guess’ process? If it’s the latter, I don’t see any good reason why this kind of best guess recognition couldn’t happen in our peripheral vision, where acuity is poor. It would make sense to have some mechanism to sort the good from the bad when things catch our eye, and even more sense for it to be context sensitive.
What makes this interesting is that it has implications for interpreting brain functions based on when they happen. Say we’re interested in the function of some neurons that are active when something catches our eye, but they are active before we have chance to focus on whatever it was, then their function should be pretty basic; it couldn’t be anything that makes use of the kind of high resolution imaging available at the centre of our vision. However, if we can identify fuzzy objects in our peripheral vision, then it widens the scope of what they could be doing.
For example, you’re at the dinner table, and there’s movement out of the corner of your eye and the neurons fire. If they’re all Mr Magoo, then they couldn’t be doing much more than telling you something has moved, and maybe it’s something you should check out. But if they can access a rough image, say a plate of round green things, they might take a guess that it’s a platter of juicy grapes, and the neurons might be able to tell you that something, which might be a platter of grapes, has appeared. There might be some value attached to that response, if you particularly like grapes. But there’s more to it than that. Firstly, the guess is exactly that – the best guess given the limited information they have. If the platter turns out to be a plate of brussel sprouts, then you’re going to be disappointed, but these neurons don’t know that. Secondly, the guess is context sensitive. If you’re at a Roman fancy dress banquet, you could safely bet your last dinarius that those green things are grapes. However, if it’s Christmas Day at Auntie Mabel’s, they’re more likely to be sprouts.
Pictured: Not grapes.
It seems obvious now, but the picture in the visual system, and probably information in most of the brain, is constantly changing based on best guesses and interpretations, and it’s not just a linear flow chart where one step completely precedes another. It’s going to make for some interesting discussion in my viva…
I’ve had a few breaks in posting before, usually because I just didn’t get round to it for a while. I’m on my first intentional hiatus at the moment. I’m finishing off my PhD, so everything that’s not working on my thesis is totally off the radar (hence why I’m up at 3am GMT doing analysis for the nth night in a row). I’ll be out of action until at least the beginning of October, and possibly longer, depending on vivas and corrections.
Mathematics is vital to neuroscience, biology, and science as a whole. Whether it is statistical examination of experimental data, or differential equations describing the behaviour of dynamical systems, maths is involved. And increasingly so: as the processing power of computers increases, new theoretical fields that examine vast swathes of data, such as bioinformatics, can open up. As subjects that are covered on introductory degree courses increasingly rely on maths, the skills taught in GCSE maths become increasingly inadequate. When students get to degree level there is often a resistance to doing maths; I’ve personally taught students who have stonewalled at the first mention of the mathematics behind statistics. While you might be able to learn how to perform a particular maths based method without understanding the calculations behind it, without understanding at least the concepts they are based on you won’t understand why you perform that particular method. Or more importantly, in which circumstances you shouldn’t perform that method, and how to tell which is which.
So do we need to teach GCSE mathematics better to our under 16s? Or make A-level maths compulsory for everyone, or at least those who intent to take a maths centred subject? Perhaps not. Part of the problem with maths is that it is so abstract. The students that I taught, after a bit of goading, understood the concepts they needed to as long as they were explained in concrete terms.
Stephen Curry hits on this in his THE article and blogpost: The way to get science students to learn and understand maths is not to teach abstract maths better, but to teach maths as an integral part of science. Although maths can be studied as an abstract field, for science it is a tool. Mathematics is a way of describing quantitative situations and problems, and applying it to concrete concepts makes it easier to understand for most people.
Note that this is not a different way to teach maths to those who don’t ‘get’ maths the ‘normal’ way. Curry points out that Fourier developed his analysis method to determine the conduction of of heat along an iron bar, and Alan Key in Doing with Images Makes Symbols relates an informal survey that found the vast majority of top mathematicians (all but a few out of 100) worked on maths problems by visualising or feeling the problem, rather than symbolic abstractions. Practice is key too: no one expects to be able to play piano with any degree of competency without practising, but people switch off early with maths, because they don’t ‘get’ it. Yet Curry reports that on his Imperial course there is no detectable effect of having a maths A-level on final results – the key is the ability to learn and practice applied maths.
Some see a problem beyond scientists using maths, and suggested that we need new ways of representing mathematical concepts in general; that abstract notation makes no sense in a world where we have the technology to create interactive visual and visceral representations. Teaching maths by using equations, the article suggests, is like teaching everyone to use computers through the command line.
No reasonable person would expect to be able to pick up most subjects immediately, or to commit to memory abstract rules and theories without having an understanding of how they related to concrete examples. Somehow, mathematics has remained at least partially separated from this reasonable expectation.
Recenly I gave a talk for Psychology in the Pub about why I think neuroscience research is particularly prone to over-simplification, misunderstanding and mischaracterisation, using the portrayal of dopamine as a “feel good drug” as an example. My slides are under the Documents section, and I’ll probably put up a summary of the talk at some point, but in the meantime there’s quite a nice article on SciAm blogs that demonstrates one of the points that I made.
In the talk, I suggested that one of the reasons neuroscience research is particularly vulnerable to misunderstanding is that we form intuitions about how the brain works based on our experience of what it’s like to be us – we will try to interpret information in a way that makes sense based on what we expect. Unfortunately, our intuitions are usually a little bit off. Four Things Most People Get Wrong About Memory gives instances where our memory doesn’t act like we think it does: we miss gorillas in our midst, remember what agrees with out expectations, forget where we heard something, and we not only trust our memories, but also mistakenly use confidence as an indicator of accuracy. If our beliefs about the way our day-to-day psychology is misguided, then attempts to interpret neuroscience research in light of these beliefs are likely to be equally misguided.
I’ve been having a bit more of a practice with Adobe Illustrator (CS2), and with the help of thesefewposts from Prof Like Substance and a more extensive guide on the MIT website, and I can now do the basics without wearing out Ctrl, Z, and my patience. So here’s the fruits of the last hour or so’s efforts (click to embiggen):
The outlines of the trigeminal nerve ganglion and brainstem trigeminal nuclei in a rat (transverse section).
A diagram of the afferent and efferent connections of the brainstem trigeminal nuclei.
1: Lemniscal pathway
2: Paraemniscal pathway
3: Extralemniscal pathway
Abbreviations: 5G: trigeminal nerve ganglion; Pr5: principal trigeminal nucleus; Sp5o: spinal trigeminal nucleus oralis; Sp5ir: spinal trigeminal nucleus interpolaris, rostral subdivision; Sp5ic: spinal trigeminal nucleus, caudal subdivision; Sp5c: spinal trigeminal nucleus caudalis; SC: superior colliculus; Pom: medial part of the posterior thalamic nuclei group; VPM: ventral posterior medial thalamic nucleus; VPMvl: venterolateral region of the ventral posterior medial thalamic nucleus.
Scicurious, formerly of Neurotopia has also moved to SciAm blogs as the The Scicurious Brain. Neurotopia looks a little quiet at the moment, if I’ve missed the other writers moving on to new pastures please let me know in the comments.
That’s not worth a pony – are you having a giraffe? Mind Hacksreports on a case study in the British Medical Journal of 26 year old male tested for language disorder for using what turned out to be authentic street slang. Probably more up to date slang than this though, me old mucker.
I’ve just been introduced to Pecha Kucha by Dave over at the Social Emotions blog. Pecha Kucha is a talk format where you get 20 slides, one slide every 20 seconds. Check out Dave’s post which involves a video of our colleague Glyn Hallam talking about emotional regulation in a talk titled Keep Calm and Carry On! How your brain can help, for the Vitae Yorkshire and North East Hub Public Engagement Competition.
Another reminder that despite how it appears, it’s probably less effort to do that work now, rather than leave it till later, from [citation needed].
I’ve been pointed at an interesting piece from an old issue of Nature looking at some terms in science with the most disputed definitions (whether or not these terms actually are the most disputed is, presumably, up for dispute).
There are some that I would consider fuzzy buzz-words like “paradigm shift” that was originally used to describe a shift in a field akin to that of astronomy moving from geocentric to heliocentric models, but is now more widely used to mean “dramatic new discovery”. If you’re worried that a term sounds grandiose, it’s probably best to save it until you’re really sure it’s deserved.
There are a couple of more what I see as ‘technical’ definitions – stem cell and epigenetics, where trendy new research terms have spread to cover areas related to, but not necessarily covered by the original definition of the term. Whether ‘race’ represents a meaningful concept in modern biology is also addressed.
Two that have more direct relevance to me are ‘significance’ and ‘consciousness’. I’m trying to get out of the habit of describing something as significant, even when it’s quite clear I’ve not applied some sort of statistical test. This is partly out of some misplaced sense of correctness, but also to reduce the chance that I’ll slip up when describing some aspect of my data. I’ve settled on “noticeably” as a reasonably replacement. Misusing consciousness is less of an issue for me and more of an interest. There’s a disjoint between using conscious in a more clinical sense to mean awake, and in a psychological sense to address…What? Some aspect of attention? Awareness? Purposeful thought? Is a choice a conscious one if you haven’t considered it?
Do you agree with their disputes? What terms are disputed in your field? Have you ever read research only to realise the author is defining their terms differently?
Inspired by the statistics classes I’m currently teaching, and a conversation I recently had in the pub with some colleagues (because I’m just that exciting), I’ve been wondering about why p < 0.05 is the most common threshold for statistical significance, at least in the psychological sciences. I realised that the choice of threshold was probably arbitrary to a certain extent, but I thought that maybe it was at least a useful arbitrary value for whatever purpose p values were first used for. I had been teaching about t-tests, so they were on my mind. I knew that the Student’s t-test was created by William Gosset to help quality control at the Guinness brewery (the brewery forced him to publish under a pseudonym – Student – to conceal from competitors that they were using statistics). Perhaps a false positive rate 1 in 20 was considered to be a reasonable error rate in brewery quality control? Apparently not…
Recent Comments