collection VOL.14
THE BIG IDEAS IN SCIENCE SI MP LY E XPL A INED
ME TRAVEL FREE WILL WORMHOLES GENE DRIVES MO OR E UROMODULATION MICROBIOME PROFILING ARTIFICIAL INTE CIAL INTELLIGENCE QUA N T UM PH Y SIC S DARK MATTER BAN FREE WILL NEUROMODULATION WORMHOLES GENE DRIVES DRIVES TIME TRAVEL DARK MATTER THE BIG BANG TIME OMODULATION SCHRÖDINGER’S CAT MICROBIOME PROFILIN TRAVEL WORMHOLES GENE DRIVES MO OR E ’ S L AW NEUROM UA N T UM PH Y SIC S ARTIFICIAL INTELLIGENCE SCHRÖDINGE XXXGENE DRIVES MO OR E ’ S L AW DARK MATTER THE BIG BA CRORBIOME PROFILING THE BIG BANG GENE DRIVES MOOR NEUROMODULATION FREE WILL SCHRÖDINGER’S CAT MICR CE QUANT UM PHYSIC S TIME TRAVEL THE BIG BANG GENE
C O N T E N 40
38
68
32
94
48 4 BBC SCIENCE FOCUS MAGAZINE COLLECTION
26
T S
N ET U C IEEENT CSE O RP OTSW
08 TUNE UP YOUR BRAIN
Memory, problem-solving, brain-training
HT EWTAEOLEPTTHS
26 DIET TRENDS
The science of fasting and microbiome tests
32 COMBATING PTSD
Deleting memories to beat the condition
38 BREAKING GENETIC LAWS
Changing gene drives to eradicate disease
40 MDMA FOR PSYCHOTHERAPY
How ecstacy could help treat alcoholism
ATNOCPI ETNWT ELEI TF SE
08
48 THE WALK OF PREHISTORIC LIFE
Robots reveal how ancient creatures moved
54 BUILDING A NEANDERTHAL BRAIN Bringing our ancient cousins to life
PTHWTY EOS EP I CT SS
62 BEFORE THE BIG BANG
What really happened 13.8 billion years ago
68 THE PROBLEM WITH GRAVITY
A new theory to rewrite the laws of physics
82 54
62 98
74 HOW TO TIME TRAVEL
From wormholes to cosmic strings
82 THE REAL SCHRÖDINGER'S CAT
New research to prove the feline could exist
TTEOCPH N O LEOE GT YS TW
90 THE END OF MOORE'S LAW
The new tech to replace silicon chips
94 DEEP LEARNING
The next big steps in artificial intelligence
98 INTERVIEW: JIM AL-KHALILI
On why we can trust artificial intelligence
BBC SCIENCE FOCUS MAGAZINE COLLECTION 5
6. DECISION-MAKING So-called free will and the emotionally fraught business of choice
Every decision we make is arrived at through hugely complex neurological processing. Although it feels as though you have a choice, the action that you ‘decide’ to take is entirely dictated by automatic neural activity. Brain imaging studies show that a person’s action can be predicted by their brain activity up to 10 seconds before they themselves become aware they are going to act. This has huge implications for our concept of free will, which scientists and philosophers are still grappling with today. Multiple neuroscientific studies show that even those important decisions that feel worked out are just as automatic as knee-jerk reactions (although more complex). The sense of volition seems to be a clever illusion perpetrated by our brains, and the illusion is useful because it gives us a sense of responsibility – and causes us to moderate our behaviour accordingly. Decision-making starts with the amygdala: a set of two almond-shaped nuclei buried deep within the brain, which generate emotion. The amygdala registers the information streaming in through our senses and responds to it in a split second, sending signals throughout the brain. These produce an urge to run, fight, freeze or grab, according to how the amygdala values various stimuli. Before we act on the amygdala’s signals, however, the information is usually processed by more sophisticated brain areas, including some that produce conscious thoughts and emotions. Areas concerned with recognition
20 BBC SCIENCE FOCUS MAGAZINE COLLECTION
work out what’s going on, t hose concerned wit h memory compare it with previous experiences, and those concerned with reasoning, judging and planning get to work on constructing various action plans. The best plan – if we are lucky – is then selected and executed. If any of this process goes wrong, we are likely to dither, or do something silly. The various stages of decision-making are marked by different types of brain activity. Fast (gamma) waves, with frequencies of 25 to 100 Hz, produce a keen awareness of the multiple factors that need to be considered to arrive at a decision. If you are trying to choose a sandwich, for instance, gamma waves generated in various cells within the brain’s ‘taste’ area compare the taste of ham, hummus, wholemeal, and so on. Although it may seem useful to be aware of the full range of choices, too much information makes decision-making more difficult, so irrelevant factors get dismissed quickly and unconsciously. So, for example, at the sandwich counter, the cheese and tomato option might trigger only the tiniest flurry of neuronal excitement. After this surge of activity marking the comparison stage, the brain switches to slowwave activity (12 to 30 Hz). This extinguishes most of the gamma activity, leaving just a single ‘hotspot’ of gamma waves which marks
NEUROSCIENCE
TUNE UP YOUR DECISION-MAKING ■ Make a list of your bad decisions and look for links. Were they all compromises, for example, or made hastily? When you identify a link, analyse the mental strategy that you used, and try deliberately using the opposite strategy for a while. If the problem seems to be haste, for example, delay the decision, and be ready to acknowledge any vaguer, subtler factors that come to mind.
GETTY IMAGES
■ Brainstorm before a decision, then sleep on it before acting. Like creative thought, good decision-making benefits from unconscious incubation, in which the brain drifts around, rummaging through memories that might be useful. Sleep is an extreme case of incubation, and your dreams may throw up important clues that make your decision clearer upon waking.
the chosen option. Although there is no ‘you’ outside your brain to direct what it’s doing, you can help it to make good decisions by placing yourself in a situation which is likely to make the process run more smoothly. Doing something that is physically or mentally stimulating before making a decision will help your brain produce the initial gamma waves that generate awareness of the competing options. Getting over-excited, on the other hand, will prevent the switch to the slow brainwaves, making it much harder to single out a choice. Subjecting yourself to high emotion may also warm up the connecting pathways from the amygdala to the action areas of your brain, causing panicky or impulsive behaviour. 5
■ Mentally step back from the situation and ask yourself what others might do. This will force your brain to look at the situation from a new perspective, which may reveal factors you had not previously taken into account. ■ Write down your favoured option, then highlight the emotional words. If you delete the highlighted words, is the decision still looking good? If not, these words are your real reasons for the decision. ‘Attractive’ or ‘exciting’ may be valid factors in deciding on a date, but not so good if you’re choosing an accountant.
BBC SCIENCE FOCUS MAGAZINE COLLECTION 21
WHAT SCIENCE HAS TO SAY ABOUT... INTERMITTENT FASTING, PERSONALISED NUTRITION & MICROBIOME PROFILING
DIET TRENDS
HEALTH
INTERMITTENT FASTING Is the 5:2 diet any better than cutting your calorie intake over the longterm? And are there any risks? words by H AY L E Y B E N N E T T
GETTY IMAGES
O
ff the back of some best-selling diet books, fasting is more popular than it’s ever been. All you have to do is keep your calories down for a couple of days each week, eat what you like the rest of the time and you’ll still lose weight. That’s the gist of the 5:2 diet anyway. But with popularity comes scrutiny, and scientists are now trying to work out whether there is any benefit to fasting diets that you can’t get from a normal calorie restriction diet – in other words, just eating less all the time. Krista Varady, who is based at the University of Illinois in Chicago, has been researching intermittent fasting as an alternative to everyday calorie restriction. “People quit those diets after about a month or two as they just get sick of the daily deprivation,” she says. “With alternate-day fasting, or 5:2, you basically get a break from dieting every other day, or five days a week.” Some of Varady’s early studies were carried out on mice. But the way she sees it, there’s only so much you can learn from animals with “perfect adherence”; if you put a mouse on a diet, there’s really not much the mouse can do about it. People, however, are another matter. “It doesn’t matter if the diets produce all these amazing results if people can’t actually stick to them,” says Varady. 5
BBC SCIENCE FOCUS MAGAZINE COLLECTION 27
A USER’S GUIDE TO
TIME TRAVEL Einstein’s General Theory of Relativity has raised another issue, which has given numerous scientists sleepless nights because it makes one thing pretty much unavoidable: time machines... words by M A R C U S C H OW N illustrations by M E R I J N H O S
74 BBC SCIENCE FOCUS MAGAZINE COLLECTION
PHYSICS
...It comes down to the fact that, in Einstein’s theory, time is not absolute,
Watch a clip from How to Build a Time Machine bbc.in/2TYdbGI
ticked off by a universal clock with which everyone agrees, but instead is relative. “I can’t talk to you in terms of time – your time and my time are different,” wrote the English novelist Graham Greene. According to Einstein, the rate at which time flows for someone depends on how fast they’re moving relative to you and the strength of the gravity they’re experiencing. If you can find a way to jump from a region where time flows at one rate to a region where it flows at a slower rate, you can go back in time – you’ll have made a time machine. The recognition that time is not what you think it is goes back to the Special Theory of Relativity that Einstein published in 1905, and it all hinges on the unique properties of the speed of light. Einstein realised that nothing can travel faster than light – it is the cosmic speed limit of our Universe. This makes light uncatchable by anything. He also discovered that intervals of space and time stretch like elastic as massive objects move through them. By a cosmic conspiracy this means that everything measures exactly the same speed for a light beam, no matter how fast that thing is travelling or in which direction. To be a little more precise, moving clocks run slow. So, if someone flies past you – and it has to be at a speed approaching 300,000 kilometres per second – then their clock will run slow compared to yours. If they could ever reach the speed of light – which is impossible for a material body, though possible for a massless entity such as a particle of light (a photon) – time would come to a complete standstill. 5
BBC SCIENCE FOCUS MAGAZINE COLLECTION 75
TECHNOLOGY
CAN WE TRUST ARTIFICIAL INTELLIGENCE? Deep learning is used in everything from speech recognition software to the assessment of mortgage applications. The only trouble is, we don’t really know how it works…
D
words by P E T E R B E N T L E Y
ILLUSTRATION: SAM FREEMAN
eep learning does it all: face recognition, language translation, game-playing. It’s an approach that has transformed the field of artificial intelligence (AI) and is the AI flavour of the decade. But how does deep learning work? And can we trust it to work for safety critical applications such as self-driving cars? We generally like computer algorithms to be as transparent as possible and this is not the case with deep learning. In essence, deep learning is a clever rebranding of an earlier computer learning method called artificial neural networks (ANNs). Dating back to the beginning of computers, ANNs are programs that simulate networks of neurons like those in our brains. They’re hugely simplified and don’t work in quite the same way that real neurons do, but nevertheless, they enable computers to learn (see ‘How a neural network works’, on page 96).
WE CANNOT SEE INSIDE THESE GIANT NEURAL NETWORKS AND THAT IS NOT A GOOD THING
HIDDEN DEPTHS Research on neural networks started in the 1950s and as the ideas were refined it soon became clear that neural networks were not as good as some other approaches in machine learning (the branch of AI dedicated to helping computers learn from data in order to make classifications and predictions). As a result,
research in the area began to wane by the early 1990s and learning methods that relied on clever statistics started to dominate. This all changed about 20 years ago. British pioneer Geoff Hinton (University of Toronto and head of Google’s Brain Team Toronto) and Jürgen Schmidhuber (IDSIA Dalle Molle Institute for AI, Switzerland) introduced new, more efficient ways to train neural networks containing far more layers. Suddenly networks could have hundreds of ‘hidden’ layers – rows of neurons that sit between the input neurons, which are connected directly to sensors, and the output neurons, which provide the results. When combined with new ways of connecting the neurons to each other, the result was massively more powerful. The breakthrough coincided with the age of big data, cloud computing and fast processors. By 2006 it was possible to create gigantic ‘deep’ networks, train them with vast quantities of data and use huge numbers of fast computers all working in sync. This ‘deep learning’ was the start of the newest revolution in AI. Although still based on a simplified model of how the brain functions, it relies on networks of thousands or millions of neurons simulated in software. Given enough data and enough computers to enable the networks to adapt and learn in response to the data, the result is like a little software brain. If it’s been trained to recognise faces, then the little brain can be placed into a camera, so that when you take a photo it finds faces and ensures they are in focus. If it’s been trained 5
BBC SCIENCE FOCUS MAGAZINE COLLECTION 95
PART OF THE
COLLECTION
THE BIG IDEAS IN SCIENCE S I M P LY E X P L A I N E D This BBC Science Focus Special Edition reveals the latest research in the fields of neuroscience, health, ancient life, physics and technology. IN THIS ISSUE… What if the Big Bang wasn’t the beginning How gene drives could eradicate disease The truth about brain training games The robots revealing how ancient life walked The black hole that could prove Einstein wrong Why scientists are growing Neanderthal brains
PLUS – subscribers to BBC Science Focus Magazine receive FREE UK postage on this special edition
5 ways to time travel – from wormholes to cosmic strings
What science says about personalised nutrition, fasting and microbiome profiling
Order online
www.buysubscriptions.com/ bigideas
ONLY
£9.99 INC. FREE UK POSTAGE*
How quantum computers, memristors and graphene processors work
Or call 03330 162 138†
and quote Big Ideas In Science Explained PRINT 1
† UK calls will cost the same as other standard fixed line numbers (starting 01 or 02) and are included as part of any inclusive or free minutes allowances (if offered by your phone tariff). Outside of free call packages, call charges from mobile phones will cost between 3p and 55p per minute. Lines are open Mon to Fri 8am – 6pm and Sat 9am – 1pm. *Subscribers to BBC Science Focus Magazine receive FREE UK POSTAGE on this special edition. Prices including postage are: £11.49 for all other UK residents, £12.99 for Europe and £13.49 for Rest of World. All orders subject to availability. Please allow up to 21 days for delivery.