Can new technologies probe human thoughts and feelings without us even realising? This week we talk to a researcher who's using mobile phones to tap into peoples' emotions to provide new insights into human behaviour and even spot the triggers that might be encouraging someone to smoke. Plus, how data mining and computer simulations can identify the patterns of behaviour that predate disasters so they can be predicted - and prevented - in future. And with the surge in online social media of the last 5 years, is statistics capable of keeping up when it comes to doing research using these resources? Meanwhile, in the news, we hear what causes cancer to spread, how ancient stone age man used bug-repellent bedding and how a Taxi driver's brain changes as he learns "the knowledge" of London's streets...
In this episode
01:38 - Gleaning the mood of the nation
Gleaning the mood of the nation
with Dr Jason Rentfrow, Cambridge University
Chris - So first of all, kick off and tell us how have social scientists collected data in the past and what are the problems with using those techniques?
Jason - Well, most research in psychology collects data in laboratory settings. So just about any psychology department at any university, you'll have a space designated for psychologists to recruit university students and then to subject these students to all sorts of interesting, often mundane, tasks. Essentially, what ends up happening is that these university students will complete a variety of different questionnaires, they may be presented with various stimuli and asked to respond to these stimulus and may be looking at reaction time, how quickly people hit space bars, and occasionally, there may be hidden cameras present in these rooms, and the researchers will then watch the videos of these participants' behaviour and code them on various indices.
Chris - This is not however, a natural environment for those individuals, so there's a real danger they're going to tell you what you want to hear or you're basically changing the outcome by the way you're measuring the data.
Jason - Well yes - that's exactly right. In theory, psychology and in particular, social psychology, the area that I work in is, interested in understanding how people behave in the natural environment and a lot of psychologists will go to great lengths to create a seemingly natural ambience or atmosphere within these laboratory settings to simulate these real world experiences. Of course, the big advantage of working in a laboratory is that you have a lot of control over extraneous variables and that's one of the main reasons why researchers will do their work in labs as opposed to the real world.
Chris - So given the constraints then, what are you trying to do to get around this problem?
Jason - Well essentially, we're trying to use new technologies to understand how people behave in their everyday lives. I was approached a few years ago to begin collaborating with some computer scientists Cecilia Mascolo who's at the computer laboratory at Cambridge, and together with other colleagues, we've started using the sensors in Smartphones and really, any of the typical Smartphones that are on the market now are capable of measuring a wide variety of different sorts of behaviour. So for example, Smartphones have accelerometers, light sensors, obviously have microphones, GPS, and by extracting data from these different sensors, we have the potential to observe how people behave in their natural environment.
Chris - Because they're more comfortable using the phone because they're well acquainted with using it, it's not an artificial situation for them. It's incorporated into their everyday lives. It doesn't change their behaviour, its involvement, and therefore, you get a more accurate representation of what's real for that person.
Jason - That's absolutely correct. I mean, I don't know the most recent statistics but millions of people carry around a Smartphone with them every day and by writing applications that will essentially extract information from these sensors and then by doing more controlled research, we can begin looking at how behavioural profiles, as measured by these sensors, may represent certain dispositions, or people who are in certain locations engaged in certain tasks which can be extremely informative from a social scientific perspective.
Chris - So what sorts of problems are you going to be focusing on? What can you use this to address?
Jason - Well currently, the work that we've been doing has looked at emotional expression and essentially, what we've done is designed an application whereby, participants will carry around their phone that's running this application as a kind of motion sense. The phone will register when the user is speaking and there is a predefined set of emotion categories in this application, and essentially, the user's speech is analysed in real time, and it looks at certain speech parameters and compares them to this predefined dictionary. Essentially, it will make inferences about the likelihood that the person may be experiencing happiness, sadness, fear, anger, or being in a neutral state. And so, this is one way in which we can look at emotional expression and experiences in a real world and then pair that information with location base information as well as whether or not the user is in a social interaction and perhaps with whom they're interacting.
Chris - That's passive acquisition of information. What about flipping the coin over so that the phone then does something back to the person? Is that something that could be done in the same sort of way?
Jason - Yes, that's another direction that we're going in collaboration with some psychologists and computer scientists at Southampton, we've recently secured funding to essentially design behavioural interventions that we can deliver using these mobile phones. Behavioural interventions getting at, for example, smoking cessation or weight management.
Chris - So you could take say, the behaviours that you know tend to be displayed by someone who might be about to relapse and have a cigarette when they're trying to quit because you know what those sorts of behaviours based on their profile would be and then you could trigger the phone to engage some kind of helping strategy or distraction or phone a friend to dissuade you from lighting up.
Jason - That's exactly right. There have been studies that have used text messaging to help people quit smoking and in these types of studies essentially, users or participants will explain and provide information about their reasons for wanting to quit, their motives, and they will provide information about the triggers that make it difficult for them to quit. By providing text messages without any other information, you may be sending messages whenever they're not really needed, but using the sensors in these phones, we might learn from the participants that on Friday nights, whenever they go to a particular pub or when they're with a certain group of individuals, their urge to smoke is strongest. So we can essentially programme this information into the user's phone and then trigger messages to them whenever they're in these vulnerable situations.
Chris - There is still a slight snake with this, that being that it relies on the fact that it assumes that everyone's got a Smartphone and there are many people - actually, I don't own one (there's a confession) so there are many people who wouldn't have one and they would lie outside the scope of the acquisition of your data. Do you think this is a problem that will be solved by time anyway?
Jason - I suspect so. I think adoption rates are quite high for these Smartphones and you're absolutely right, that there are constraints. However, even other phones have capabilities that would still allow for delivering such interventions. They just may not have all the full functionality as on the Smartphones.
Chris - But what I'm getting at is that a lot of the behavioural monitoring you're talking about doing, is there not a risk that the kind of person who will have a phone and the kind of person who will sign up to let you eavesdrop on their behaviour may not be representative of the population at large so you've still got a problem?
Jason - Yes, that's absolutely right. I mean, not everyone as you said has a Smartphone and not everyone would be comfortable using such technology even if they did have a Smartphone, and that certainly will be a limitation in the research that we're doing, we are using other methodologies. Some of our collaborators have been very successful at delivering these interventions online - so for desktop computers or laptops for example - and will be doing other studies in more controlled settings to essentially not only see whether or not there are user differences depending on which form of the intervention they choose but also, whether certain interventions may be more effective than others.
Chris - Jason, thank you. That's Jason Rentfrow from Cambridge University.
10:18 - Finding the Hidden Rules of Society
Finding the Hidden Rules of Society
with Professor Steven Bishop, UCL
We've heard about how we can collect data, but if we collect enough data about our behaviour, we can start to make models that can begin to reveal the hidden rules of how we behave. These models could then be used to predict important events, like changes in economic markets, and help us to work out how to react to them. And we're joined this evening by Steven Bishop from the Department of Mathematics at University College London and he's helping to develop just such a system. It's a project of FuturICT, so thanks for joining us, Steven.
Steven - Pleasure.
Kat - What I'm really interested in is this "interconnectedness of systems" and we've seen in the recent financial problems the world had that some dodgy mortgage deals in America have affected my parents' pension. So how interconnected is the world? How has this happened?
Steven - Things have become very interconnected and in fact, so interconnected that we don't quite understand how they behave and respond. So it's not just the fact that they are interconnected, that itself doesn't lead to problems, but it's the behaviour that induces somehow. So we need to understand the hidden effects of the interconnectivity, that hidden behaviour.
Kat - One of the problems with this is that we're in a situation we've never ever been in before and with many types of science, you can observe things over time and make predictions and do experiments. We're in an ever changing world. How do we gather data to help us understand and make models?
Steven - Well the data would come from a variety of sources. We we're just talking about survey data or people in a laboratory, but there's all sorts of data from the sensors of mobile devices, but also, there are plenty of citizen scientists who actually want to give their data to make better projects and to have better data. So the whole citizen science thing is becoming much bigger. It's a collective behaviour.
Kat - You're involved in a project called FuturICT and one of the things you talked about is the idea of a Planetary Nervous System. What's that and how does that work?
Steven - We believe that there's so much data available now that almost everything is out there, and it's just a matter of trying to find it. Firstly, to get hold of that data but secondly, trying to understand it, really understand what it means. We call it reality mining. So, we really understand what the real implications are of that data. It may come from fixed sources or mobile devices, or indeed media even. We do lots of searching of media to try and understand trends in the media.
Kat - So to make a little comparison, we've discussed previously on the show, and a lot of people know about, how we model climate systems. We take a lot of data, and they make models of climate systems to see how the climate might change in the future. What sort of outcomes do you think you can predict with the kind of modelling and data gathering you're doing and what will people do with these models and the outcomes?
Steven - Well there are three places I think that the models could be useful. Firstly, for policy makers, they could actually help us to make better decisions. More evidence based policy could be done so it stops us doing this pinball policy situation that we have tended to go through in the past. So, that's the first aim. The second aim is for businesses who could really use data and models in different ways, and the FuturICT project would be really a platform for them to do business from. Rather like Amazon now - it started out as book seller but now, it's actually a platform for many businesses to interact with. But also thirdly, for the individual. Individuals and groups could decide their collective behaviour or how they could do things collectively which nonetheless had a global implication to it. Those are the three areas that I think could be very useful.
Kat - I know that part of the project involves gathering data from mobile phones and social networking and Jason has talked a little about the kind of things that mobile phones can do. But some people are getting very concerned about privacy issues, there has been some discussion about Facebook invading people's privacy. How can you get people to give up data without invading their privacy or do people just give you all their data or don't really care?
Steven - Sort of both, really! Firstly, there are many people who are quite happy to give their data in these citizens science type environments. Some colleagues from LSE did a project at the Lord Mayor's Show where people gave up their data to say where they were. In return, they got back which bars were the best to go to with the smallest queue. So sometimes, people give up data and get something in return for it. It's not just a one way thing. It should be that it goes back the other way as well.
Kat - So this idea of constantly gathering data, building models, finding out where we are and what's happening with all these really complicated systems, I can see that that's going to be very useful in the future, but are there any examples of where this kind of technique is going on right now?
Steven - Well in the UK, we did lots of studies, not me personally, but people trying to understand how the spread of foot and mouth disease occurred. We also did things like the H1N1 spreading so colleagues of mine in the US did very big studies of that. In these problems, we can see the benefits of doing scenarios - "what if?" scenarios - what would happen if we close the airports down? We can actually run "what if?" scenarios much better on the computer than we can in real life. This is what we should be doing I think.
Kat - And I guess the idea would be to avert complete disasters of the future by seeing where things are starting to go wrong and pulling things back from the edge before they get too bad.
Steven - That's correct. Coming back to your starting phrase about connectivity, we've seen several cases where connectivity causes cascades of failures, knock on effects that we sometimes might have guessed if we'd known about them, but other times, they give really non-linear effects which we really cannot understand how they would knock on to other systems.
Kat - We are all connected. Thanks very much. That's Professor Steven Bishop from UCL.
16:33 - Seed and soil – understanding how cancers spread
Seed and soil – understanding how cancers spread
Cancer is a growing problem - and because we're all living longer, we're more likely to develop the disease at some point in our lifetime. Although we've got a lot better at treating the disease in recent years, it does become a lot more tricky once it has spread through the body - a process called metastasis, which is the main cause of 9 out of 10 deaths from cancer.
So there's understandably a lot of interest in understanding why cancer spreads, and how we can stop it, and in a paper in this week's issue of the journal Nature, Ilaria Malanchi and her colleagues have made an important step forward.
Cancer spreads when cancer cells break away from a tumour - the primary tumour - and travel around the bloodstream or lymphatic system, looking for new places to start growing, forming secondary tumours. The researchers were intrigued by the fact that although many cancer cells set off around the body from a primary tumour, only a very small number of them actually form secondary tumours. It's a bit like a dandelion throwing off thousands of tiny seeds, but only a tiny handful of them actually landing somewhere where they can grow.
In this new paper, the researchers have discovered that it's a bit of both. The scientists were studying an animal model of breast cancer that spreads to the lungs, and found that secondary tumours in the lung could only be started by a small group of cancer stem cells - these are the 'immortal' cells thought to be at the heart of many cancers, and make up less than 5 per cent of the cells coming from a tumour. So, there's definitely something special about these cells. But still, not all the available cancer stem cells went on to form secondary tumours, so there must be something special about the places they land too.
The scientists homed in on a protein called periostin, which is produced by supporting cells - known as stromal cells - found in the places where healthy stem cells grow, and also in places where secondary tumours grow. The scientists found periostin in the stromal cells in secondary tumours in the lungs, but when they looked at normal, healthy lung tissue, they didn't see any periostin -and, importantly, they didn't see periostin in the lungs of mice who had breast cancer that hadn't yet started spreading.
When the scientists looked at mice with cancer that had been genetically engineered to lack periostin, they found a massive reduction in the number of secondary tumours, proving that it's extremely important for helping cancers to spread.
The scientists think that the cancer stem cells give out some kind of signal that makes supporting stromal cells produce periostin, which in turn causes the cancer cells to switch on processes that make them settle down and grow. Effectively, the scientists think that the cancer stem cells are wandering around the body, shouting out a message to stromal cells. Some of the stromal cells hear this message and respond, making a suitable environment for the cancer stem cells to bed down and take root.
This early research is very exciting because the processes the team have uncovered could be at the heart of many different types of cancer that spread.
20:20 - Fred Flintstone's bed uncovered
Fred Flintstone's bed uncovered
Writing in this week's Science, University of Witwatersrand palaeontologist Professor Lyn Wadley and her colleagues describe an excavation they have carried out in a cave site called Sibudu in South Africa's KwaZulu Natal Province.
Dating from 77,000 years ago, the team have uncovered successive layers of sedge, and other plant materials including grasses, arranged on the floor of the cave and covering an area between 1 and three metres across. Also within the oldest deposits are thin layers of leaves from the Cryptocarya woodii - also known as the Cape Laurel - tree.
Trees of this species are well known to practitioners of traditional medicine, and chemical analysis of the leaves has confirmed that they contain a range of insecticidal compounds, including alpha-pyrones.
It's likely, therefore, that the ancient middle stone age inhabitants of this shelter were aware of the beneficial mosquito-repelling qualities of these leaves and protected themselves by using them within their bedding.
Moreover, these early humans were also pioneers of infection control, it seems, because from about 73,000 years ago there's evidence in the cave that, rather than make their beds, the inhabitants regularly burned them. This would have had the effect of helping to rid the environment of parasites and other insect pests.
The finding is extremely important because, whilst there is robust evidence of the activities of stone age peoples out in the field, their domestic arrangements were much less well understood. Maybe they even pre-empted the inability of the teenager to make a bed, which is partly why they torched theirs...
23:01 - Reducing Cancer Risk Factors
Reducing Cancer Risk Factors
The biggest study so far into cancer risk factors confirms our understanding of the lifestyle choices that may lead to cancer...
Kat - For my day job I work for Cancer Research UK and we put out a very big story this week looking at risk factors for cancer. [The paper showed] that around 100,000 cancers a year are caused by just four preventable risk factors: Smoking, drinking too much alcohol, being overweight and not eating enough fruit and veg.
Chris - But that's not rocket science, is it? Because we did know that...
Kat - It's absolutely not. But this is the biggest paper so far to have come out that's looked at lots of different types of cancer and lots of different risk factors, and actually really been able to quantify it. It's the most reliable data we have on the links between a lot of different risk factors, not just the big ones like smoking. Smoking is by far and away the biggest risk factor for cancer, but [this paper looked at] many other ones, things like UV, occupational exposure and all these kinds of things. And we've been able to put some hard and fast numbers onto these, about how much [these avoidable factors] can increase your risk of cancer.
Chris - So, can you just very briefly give us a couple of them and explain why this adds new value and new insight into those areas?
Kat - Well in some ways, it's not surprising to a lot of us who work in cancer but obviously, the really big one is showing that almost a quarter of cancers in men are down to smoking. That's a really, really big risk factor. But interestingly as well, one of the things that came out is that not eating enough fruit and veg is a much bigger risk factor in men than in women. This is because generally, men don't eat enough fruit and veg compared to women. So there were some interesting things that came out, but it was more or less a no-brainer! It's nice to see all the data there in a way that can really be explained and shown to people.
Chris - And if people would like to follow up and see this data, where should they go?
Kat - There's a really nice run down of it and a really good infographic on the
Cancer Research UK Science Update blog, so you can go and have a look at some of the factors there. There's a very interesting discussion going on there because some patients have felt that putting this information out is blaming them for getting cancer and of course, for an individual cancer, it's very difficult to say exactly what caused it. But it's more important that we can tell people we know what can increase your risk and what can reduce your risk as well.
25:15 - Navigating a Taxi Driver's Brain
Navigating a Taxi Driver's Brain
with Professor Eleanor Maguire, University College London
Qualified London taxi drivers know their way around over 25,000 streets in the capital. And, if you scan their brains, you find that the structure called the hippocampus, which contains a mental map of the world around us, is much bigger than it is in the average non-taxi driver. But was it bigger to begin with, or did learning London like the backs of their hands trigger the cabbies brains to change? Now, UCL's Eleanor Maguire thinks she knows...
Eleanor - Animals who do a lot of navigating often have a bigger hippocampus than animals of the same species who don't engage in much navigation. So we wondered if the same would be true of the human brain and whether those who navigated a lot would also have a bigger hippocampus than those who didn't navigate so much. And so, about 11 years ago, I studied this using Magnetic Resonance Imaging on some London taxi drivers and we indeed found that they had greater grey matter volume in part of their hippocampus than people who didn't navigate so much.
Chris - I guess one problem though or one criticism of that is that it's purely observational in the sense that you look at this group, they're taxi drivers. Do they have a very big hippocampus because they're taxi drivers or do they have a job as a taxi driver because they have a very big hippocampus which means they're endowed with the very good map in their head.
Eleanor - Yes and that's a very important point and in fact, that's one of the prime observations of the current study, was to try to see if we could document within specific individuals the change that might occur in the structure of the hippocampus, purely as a consequence of acquiring this very detailed mental map of London.
Chris - So how did you actually do it?
Eleanor - Well what we did was with the cooperation of the Public Carriage Office, we recruited trainees who were just starting there and training as London taxi drivers and we scanned their brains and we tested their memory. And then off they went to try to acquire the knowledge and this takes about 4 years on average, 3 to 4 years. And so, when people had qualified, we invited them back and we scanned their brains again and tested their memories again. And we were able to, in the first instance look at people before they started and see if there was anything in their brain or their behaviour that could predict who would eventually qualify because the interesting thing is only 50% of the trainees actually went on to qualify. It's an extremely tough thing trying to become a taxi driver in London.
Chris - The odds are slightly better in medicine. Did they give you any reason why they dropped out, the ones that did drop out because I mean, there may have been perfectly sound reasons other than cognitive ones?
Eleanor - Absolutely and I think it's difficult to know. It's probably quite a heterogeneous group in the sense that some people probably did find it very tough going and they just didn't have the navigational skills to pursue this, but it's also the case that embarking on this training can be time consuming, can take time away from your family, it's a big financial commitment, and in the current climate undoubtedly, some individuals had to withdraw as a consequence of those sort of issues. So it's not easy to know exactly why people dropped out and sometimes people can say they dropped out for one reason but maybe it was another reason and so on. So it is quite a mixed group, but we did end up with a group that didn't qualify, a group that did qualify and then of course, we had control participants who didn't engage in any training at all but still, we scanned at the start and at the end of the study, just like the trainees.
Chris - Probably, the most important question is, those people who you scanned at baseline and then they became qualified taxi drivers, did you see any differences in their brains?
Eleanor - We did. For those who qualified, we found that between the start and the finish of the study, the back part of their hippocampus had increased in volume and no other part of the brain had changed, just very specifically this back part of the hippocampus which is what we found in our previous sort of observational studies where we compared taxi drivers to non-taxi drivers. So it was fully in line with our previous results.
Chris - And the controls, they didn't show any changes?
Eleanor - No, the controls and those who didn't qualify, their brains remained exactly the same from start to the finish of the study.
Chris - Now what about other measures of cognition because you said you also tested their memories in other ways? So rather than just looking at the structure of the brain, you also looked at function. What were the differences then before and after?
Eleanor - Well obviously, the first thing we did was we tested people's general intellectual ability just to make sure that there were no differences in that regard. So the IQs for example of the individuals were all very similar. We then tested their basic knowledge of London in terms of understanding spatial relationships between landmarks in London. And then we did a whole range of other memory tests that looked at their ability to remember verbal material, words or pictures or other types of spatial information. So we did those tests at the start and then we did parallel versions of those tests at the end of the study.
Chris - And how did the results of the before and after compare amongst all the groups?
Eleanor - Well we found that the controls didn't change and we found that the trainees, particularly the qualified trainees became much better in terms of their knowledge of London and the proximity of landmarks to each other which of course you'd expect because they were trying to actually learn that information. But what was most interesting was that on other tests of spatial memory, those who qualified actually performed worse at the end of the study than they did at the start. And this is something we found previously in our studies of taxi drivers that although they are experts in terms of navigating around London, perhaps there's a little bit of a price to pay for that expertise in that they become a little bit worse at dealing with information of other kinds. And that kind of makes sense you know, somethings got to give when you're taking in a lot of information.
Chris - And of course, you're left with another problem which perhaps you'll answer in another 10 years which is, those people that didn't drop out and did show this change, is there something special about them and that their brain is more adaptable, it can incorporate new cells, make more grey matter when they need to do a task like this, compared with people who find it less easy?
Eleanor - Yes, I think that is another important question and so, we must consider the reasons for why people fail to qualify. It may be that there are genetic predispositions to hippocampal plasticity in the individuals who qualified, allowing them to expand their knowledge and so, expand the volume of their posterior, the back part of their hippocampus. And there may be other individual differences that come into play as well. So I think this is a very important issue because what we all want to know is given any individual, what can they hope to achieve? How much can they learn? What capacity does their memory and their hippocampus have? So I think it's going to be very important to understand these individual differences in future studies.
Chris - That's was Eleanor Maguire from the University College London and she published that work this week with her colleague Katherine Woollett in this month's Current Biology.
33:30 - Bed Bugs, Night Shifts and Deterring the Affections of Fish...
Bed Bugs, Night Shifts and Deterring the Affections of Fish...
with Coby Schal, North Carolina State Universitys; An Pan, Harvard University; Safi Darden, Exeter University
Bed Bugs from Abroad
The recent resurgence in bed bug infestations taking place in western countries is owing to the biting insects being re-imported from abroad rather than reoccuring locally. North Carolina State University's Coby Schal analysed the genetic diversity of bed bugs collected from outbreaks along route 95 on the US East Coast.
Coby - Because we found this extremely high genetic diversity in populations along the east coast, that should suggest to us that there are multiple introductions of bed bugs coming into the United States for multiple sources and that sort of pattern would argue against local resurgence of bed bugs. It suggests that they're coming from other places. It's difficult to place the blame on any one group, but I think international globalisation and commerce, and increased transport are very likely involved.
Preliminary research on genetic diversity in bedbug populations presented in Philadelphia, at the
annual meeting of the American Society of Tropical Medicine and Hygiene, December, 2011.
Increased Risk of Diabetes on the Night Shift
Working rotating night shifts can increase a woman's risk of developing type 2 diabetes by up to 60%. This appears to be at least partly down to night work also causing weight gain. Announcing the results in the current edition of PLoS Medicine, Harvard scientist An Pan looked at data from 170,000 women, aged 25 to 67, who had been followed up for between 18 and 20 years...
An - Compared to women who do not do any rotating shift works, women who have already done 1 to 2 years of shift work has about 5% increased risk. For women who do 3 to 9 years shift of work, the increase rate jumped to about 20% increase risk and they will go higher, more than 20 years of rotating shift work the increase rates go about 60% increase. Body weight explains a lot of associations people who do not shift work gain more weight during the follow up, compared to women who do not do shift works.
Rotating Night Shift Work and Risk of Type 2 Diabetes: Two Prospective Cohort Studies in Women
The Fishy Way to Deter unwanted attention
Non-receptive female fish resort to hanging out with a more attractive counterpart to divert unrequited male mating advances away from themselves. Working with guppies, Exeter biologist Safi Darden found that, given a choice, females for whom the time wasn't right would actively seek out a fitter female for company.
Safi - Now that we knew that females would receive less attention if they were with a more sexually attractive female, did they actually actively make this choice to spend time with a more sexually attractive female to avoid male attention? And when we gave females that weren't receptive to male attention, the choice to swim next to a more attractive female or an equally attractive female, we found that she preferred to swim in close proximity to the more attractive female. When we test that fish that were receptive, they didn't have any such preference. And this has important implications for how these and other fish organise their social hierarchies. It was published this week in the journal Proceedings of the Royal Society B.
Social preferences based on sexual attractiveness: a female strategy to reduce male sexual attention
36:59 - Ozone Depletion and Behavioural Change - Planet Earth Online
Ozone Depletion and Behavioural Change - Planet Earth Online
with Jonathan Shanklin, British Antarctic Survey
In May 1985 scientists discovered the hole in the ozone layer. Two years later, governments around the world signed up to the MontrealProtocol to phase out the use of the chlorofluorocarbons - CFCs - thatdamage ozone. So why can't a similar thing happen when it comes to climate change? Planet Earth's Richard Hollingham went to speak with Jonathan Shanklin, one of the original ozone hole discoverers...
Jonathan - We've got the data from Antarctica which I'd plotted up and one of the key things was that we could see something systematic going on in our data.
Richard - So, let's look at this graph - it's plotted by hand, a time before computers would do this.
Jonathan - Drawn on paper with pencil and a ruler to draw the best fit line to the data, but the point is anybody can see that something is happening. You can see ever so clearly that ozone amounts are going down and that was really absolutely key, once you had something systematic then something must be causing it, and my question was what?
Richard - This is the Nature paper we've also got here, and this is interesting because figure 2 of the Nature paper again, obviously, originally draw by hand then printed, you made a correlation between this decline in ozone and CFCs?
Jonathan - Yes. There had been thoughts that CFCs could affect the ozone layer for some time but the predictions were that they should be affecting the ozone layer high above the tropics. What we actually found was low above the Antarctic, quite a different place, but because there's been this expectation that chlorofluorocarbons could affect the ozone layer we thought, well this is probably the case and so when we plotted the graph we guided the eye by choosing the right scale so that at a glance you could see that there was a correlation between ozone amounts declining and the CFC amounts going up.
Richard - Okay, so the Montreal Protocol,a huge success in terms of environmental treaties, just getting that done two years after your discovery. Why then is there still an ozone hole over the Antarctic and now, what many people are describing as an ozone hole over the Arctic as well?
Jonathan - The Montreal Protocol has been incredibly successful. Today every single one of UN member states have signed up to it, but these CFCs are very very stable. They persist in the atmosphere a long time and although it's very clear that we've passed the maximum in the atmosphere, the amounts are going down, the treaty is working there is still so much around that ozone destruction can take place whenever the conditions are suitable. Now, in the Antarctic they're suitable every year - this year we've had one of our deepest and largest ozone holes actually on record. Also the conditions were suitable in the Arctic this year.
Richard - So all this points to much greater complexity, Jonathan, than perhaps you first imagined when you plotted this very simple graph.
Jonathan - It is a lesson in how quickly we can change our atmosphere. This happened in the space of about a decade between it being detectable to it being a full blown ozone hole.
Richard - Let's go back then to this paper, this graph and the Montreal Protocol which is given as an example a huge success when it comes to these sorts of treaties. Is that repeatable when it comes to global warming, to climate change?
Jonathan - There are differences. With the CFCs, just about everybody was on side, the manufacturers were quite happy to switch to a different product, it was very easy to switch to a different product. Also, the public don't like holes, so calling it an ozone hole that must be bad just because of the name and also the link between increased ultraviolet light and cancer. Again, cancer is one of our real bains of today's society, so if something is causing cancer we've got to get rid of it. So everything worked in favour of doing something about the ozone hole.
With the greenhouse gases it is much harder - greenhouse warming sounds nice, but it will take a very big change in lifestyle for individuals to reduce their dependence upon substances that get converted into carbon dioxide and also the manufacturers, the industry, the oil industry in particular is rather reluctant to stop selling oil. There's no cheap alternative that could be widely sold. I think we will be lucky to get a treaty that's effective like the Montreal Protocol was.
42:43 - The Death of Biostatistics?
The Death of Biostatistics?
with Arnoldo Frigessi, University of Oslo in Norway
Chris - Almost every aspect of science relies on statistics to decide whether results are valid or not, but when statistics were invented, there was no such thing as social media - Twitter and Facebook didn't exist. Now, given that people are increasingly using this sort of crowd-sourced data to do their research, are old style stats still up to the job? Professor Arnoldo Frigessi is the Director of Statistics for Innovation at the University of Oslo in Norway and he's got a perspective on this...
Arnoldo - That's true, so old statistics is dead. Two big things are happening. One is the data basis is changing dramatically; Before, we had these nicely built up case control studies, taking people that had the disease, people that didn't have the disease and comparing them carefully, maybe 20 in one group, 55 in the other one. Now we have Twitter and Facebook. Now we have millions of people out there that tell us about their diseases and about their bio facts, and their symptoms and we have to use this data to find out, as we did before in case controlled studies, if a drug works or not.
Chris - With that level of user engagement, where before we had small trials and we had to make sure we controlled everything very, very tightly, now we can do effectively massive trials. The noise is actually enormous, yet you can still extract very meaningful data of very high statistical significance from that because of the size of the sample.
Arnoldo - Exactly. So we're trading sample sizes, accuracy, or bias and we have to correct things because of course, not everybody is using the internet, right? And we have to correct to get the right population, and this we can do, that's not difficult. But today, we can monitor Google for example and find out where the flu is, there are papers out there that show that by looking at how many people, and where are the people that are checking for symptoms of flu, we can predict the wave of the flu in the world, two weeks before the World Health Organisation does it. Or we can, for example, use the happiness of people to find out if stock values are going up or down.
Chris - Just like a paper in Science that came out in the last couple of months using Twitter messages to look at dysphoria and happiness, and how this evolves as the day goes on. We're all happier in the morning, miserable by evening, except for one little surge before bed and it's true regardless of creed and culture. I was amazed by that and the fact that you can do these sorts of experiments now with high statistical significance and you don't actually have to leave your lab or recruit people. You just look at data that's in the public domain very easily.
Arnoldo - This is really completely changing how statistics have to work. Now, we have to use enormous data sets which are very bad, where we need models to filter out the noise. We need to correct biases. We need to find out if things are independent. Of course, if we check out how many people like Silvio Berlusconi today...
Chris - Well he likes himself presumably...
Arnoldo - ...Yes, exactly, but not many others! So you need to find people that are a bit independent from each other, right? So think that Facebook is an enormous network where you and your friends are connected by edges and now we have to find independent people. We can't take you and your friends. You mean mostly the same so that's very boring. We have to essentially split up this network in independent clusters. This is a completely new story. We have to sample networks to find independent units so that we can, in my words, reduce variances.
Chris - But how is statistics responding to this? How are people trying to control for this new domain that we find ourselves in, this new regime of doing research?
Arnoldo - I think this is coming. It's not yet a daily thing that we do, no. It's difficult. I mean, Facebook, they're not so happy to give you all the data, right? So there is much to do here still to get this power out there. But for example, if you move slightly, let's take an easy situation. Let's take for example a financial institution that has millions of credit cards out there, and they're checking, trying to find out if there are frauds. So suddenly, they also have enormous quantity of data, millions of clients, each of them with a little data, not very much, and we have to find out which of these credit cards are doing something strange. So now we, statisticians again in this situation, you're trying to find surprises in millions of possibilities, millions of tests, in some sense and we're looking for surprises.
Again, [this is] a dramatic change to what was before. Before, you had your own nice hypothesis. You had your gene that you like very much and you would spend all your life studying that gene. Now we have millions of genes and you're checking them all to find out if there's anything interesting. So the scientist doesn't have any hypothesis anymore. They say, "Oh! I have data. Here are my data. Find something useful please." So suddenly, the statistician is doing a completely different role. We have to find things. Not just check if they're true or not.
Chris - That's the genome-wide association study where it's "changes in search of a disease" now, rather than "[we have] a disease, now let's find the genetic underpinnings"?
Arnoldo - That's right. So it's really changing the way we are building up statistics as a mathematical instrument because statistic is mathematics, right? And so, we have to compute things that are called P values, probabilities that a certain gene is in some sense important for you to describe your disease. And now, we are producing lists of candidates, lists of hypotheses that may be useful. If I give [my biologist colleague] my list of 100 genes that I found after a careful statistical study, I can't tell him that these 100 genes are really all important but I can tell him, by using something called the false discovery rate, that 20% of them are wrong.
Chris - We don't know which 20.
Arnoldo - No, we don't know which 20, but that's alright. He can cope with that. I mean, he came with only data and I gave him quite an interesting answer.
Chris - So what's the next step then? Will we see bio informatics being taught very differently at university? When I was doing biometry at medical school, we were taught about the T-test, students T-test and Gaussian distributions, and we were taught "this is the circumstance when you apply this test". It sounds to me that we're going to have to go back to kindergarten, statistically, for researchers basically to understand how to use the tools we have in this whole new research setting.
Arnoldo - You're right. Statistics is more difficult now and statistics is more present in the core of medical science. So before, it was something that came at the end of your laboratory work and when everything was done, you had your nice Excel file, your nice table, then you would do your statistics. Now statistics comes from the beginning as a core instrument of discovery and this is more difficult so we have to do partial differential equations and networks, and all this stuff that is a bit more difficult.
Chris - It's good for you though, isn't it because it means you're not going to be out of a job.
Arnoldo - No.
Chris - It just means that researchers like me are going to find it much harder to...
Arnoldo - You have to find a statistician.
Chris - Well we have to find a statistician, but also, when we try and read papers, picking out whether or not the numbers are meaningful and it's solid, is going to be much more difficult. I guess most papers will have to float pass someone like you for you to unpick them and decide whether or not what they're saying is really a valid conclusion.
Arnoldo - I think in this world where competition is very strong, where you need that little advantage to make progress, we're looking to second order effects, to small things, 1%, 2%, and these things are much more difficult to find. In a different world, in industry for example, [they] use statistics also to make progress of course, they need to find those small differences or small incremental advantages to beat the competition. And now, you need statistics at a much more refined level. You need to extract much information from your data that is hidden; interactions, dependencies and things that happen together and therefore give an advantage.
How does a computer see meaningful trends?
Steven - Pattern recognition and data mining to reveal trends in data is a big part of computer science today. So we're learning more and more about that.Chris - The guys who did the work on Twitter and Facebook, they were using 2 billion tweets or something to work out what kind of mood people are in and what time of the day. So you're just telling the computer to look for certain words and certain words mean a good mood, certain words indicate on average a bad mood and you can extrapolate from that. The computer isn't actually reading. It's basically looking for trends which you tell it to look for.
How do you analyse data without presupposition?
Steven - Actually, different people can look at the same data set and try and find different information from it. So, it's not necessarily one approach. Different people will try and tease out different bits from the same data source.
52:41 - Is emotion detection via speech more meaningful than handwriting analysis?
Is emotion detection via speech more meaningful than handwriting analysis?
Jason - In the case of detecting emotions from speech, this is a very well established line of research and we can ask individuals how they're feeling and we can use other sorts of methods to assess emotions. Essentially, what we are doing are identifying speech cues that are valid indicators of people's psychological states. And so, in that sense, I think it's a perfectly valid and robust method.
53:23 - Can people with Autism benefit from emotion detection technology?
Can people with Autism benefit from emotion detection technology?
Jason - Absolutely. In fact, the computer scientists that I've been working with have shown that these phones are very powerful and can analyse this information in real time on the phones themselves or to send them onto a server and bring it back. Really, one of the big limitations with doing these analyses on the phone are the batteries of the phones.
Chris - There was a report from MIT, the media lab there, in the last five years or so, they had made a pair of spectacles that would use a camera to look at expressions on the faces of people that a person with an autistic disorder was talking with and could vibrate in the person's pocket if the facial expression they were showing showed that they were reacting negatively to things they were saying.
Jason - Yes and certainly, I think that this is just looking at another source of data looking at speech and it could do the same thing.
54:37 - Do fish enjoy reproducing?
Do fish enjoy reproducing?
We posed this question to Mark Bretha, Marine Biologist at Plymouth University...
Mark - The fundamental question here is whether fish are ever capable of experiencing pleasure at all, including at the spawning event when they release their gametes. One problem that we have is that we can't exactly go up to a fish and ask, "how was it for you darling?" And this is a general problem in understanding what feelings non-human animals might or might not experience. One area that has been the subject of a large amount of research in fish is the experience, not of pleasure, but a pain. In an experiment with rainbow trout, we see an injection that would've been painful to humans, they show behaviours like rubbing the affected area which went beyond a simple reflex response and that were also specific nerve fibres that responded to the injection. Therefore, fish might experience something akin to pain in humans.
As far as pleasure goes, there are some anecdotal evidence that when client fish interact with cleaner wrasse, they might enjoy the touch sensations of being cleaned.
In the case of spawning, we know about the hormonal control of the events but we don't yet know whether it's an ecstatic experience, altohugh it's perhaps nice to think that there's the possibility that cyclids can get their kicks and brill get a thrill from spawning. Diana - It's possible that there is some sort of neurological reward for fish when they reproduce, but we don't yet know if it's pleasurable or if it makes the, er, water move.