3

The vehicle and the driver

Dr Katy O’Neill, a gentle woman with a warm, wrinkled face and the softly spoken manner of a grandmother, sees the human impact of Australia’s poker machine problem every day. She is a counsellor at St Vincent’s Gambling Treatment clinic in Sydney; it is her job to try to guide those like Doug out of a gambling addiction. Through her door come hospitality workers, tradesmen, academics, mathematicians, bankers, journalists, and pensioners. ‘I’ve seen young men who haven’t eaten for two days,’ she tells me as we sit in one of the clinic’s counselling rooms one afternoon in May 2015.

The room we’re in is the main one she uses when counselling patients, virtually all of whom are in trouble because of poker machines. ‘I’ve never had to treat someone for buying too many lottery tickets,’ she says. Much effort has gone into making the room look and feel as far from clinical as possible. It is warmly lit, and has an armchair that Dr O’Neill sits in, and two plush couches that help to give it the appearance of a homely reading room. On one of the walls is a large framed print of John William Waterhouse’s 1891 painting Ulysses and the Sirens. The mythical story it depicts — of Ulysses tied to the mast of his ship, and his crew with their ears stuffed with wax to resist the enchanting but deadly call of the sirens — is one of Dr O’Neill’s favourites. ‘I think it’s an amazing story,’ she explains. ‘I tell all my clients about it because there’s such a mythology around the pokies that they are entertainment and people have free choice to play them. But I think once you’ve started to play and had a win, there’s a lot of reasons why you don’t have free choice to play.’

Dr O’Neill does not mince her words when talking about poker machines. She believes they are ‘immoral’ because, unlike with other forms of gambling, ‘the gambler hasn’t been fully informed about the game they’re playing; about the design of the game’. She also believes — given they are so widely accessible and harmful, state-sanctioned, and embedded in everyday life — that ‘Australia has pokies the way America has guns.’

Prior to starting at St Vincent’s in 2006, Dr O’Neill worked for a number of years in drug rehabilitation. Many of her patients back then arrived with at least a basic understanding of how they had come to be addicted, because of the high level of public awareness about the addictiveness of certain drugs. This, however, is not the case with most of her current patients, who express utter bewilderment at being so compelled to gamble and their inability to stop. ‘They will often say, “How did I get so hooked? Why didn’t I leave earlier? Why did I spend $1,000 when I said I was only going to spend $100?”’ Dr O’Neill explains. ‘They think they’re stupid.’

The general public thinks much the same about Dr O’Neill’s clients. A September 2015 study by the Centre for Gambling Research at Southern Cross University involving a survey of 2,000 Victorians found that gambling addicts are commonly stereotyped as impulsive, irresponsible, greedy, irrational, anti-social, untrustworthy, unproductive, and foolish.

A quick look at media commentary reveals as much. Take this from ex-AFL star Campbell Brown, who was a guest on Karl Stefanovic’s panel show, The Verdict, in October 2015 when talk turned to gambling addicts: ‘I’m sick of their victim mentality. No one’s holding a gun to their head.’ Or this from talkback radio host Tom Elliott, in a September 2016 column in the Herald Sun:

No one rounds up people at gunpoint and forces them to play the pokies. Lose too much money to the pokies at your local pub or club? That’s what gambling is designed to do — separate you from your cash. Forget blaming the intoxicating lights, loud noises and spinning symbols for your misfortune. Instead, acknowledge you hoped to win something for doing nothing — and next time just give the poker machines a miss.

Dr O’Neill thinks very differently. She understands exactly how addiction — especially gambling addiction — develops. In fact, it is something she is ‘obsessed’ about. ‘One of the criteria for if you have a gambling problem is whether you become obsessed with gambling. And, tick, I’ve got the first criteria,’ she jokes. The science behind gambling addiction is, she says, ‘morbidly fascinating’.

In order to reduce the embarrassment, shame, and self-loathing that many of her patients feel, and to help them take their first steps in recovery, one of Dr O’Neill’s main priorities in treatment is educating them about this science. ‘I’m telling them about why they’re going against their better judgement, and how there is a difference between wanting and liking something,’ she tells me. ‘I’m explaining how the pokies manipulate a user’s brain, and how their design and technology greatly contributes to addiction.’

Dr O’Neill also makes her patients aware of a recent major development in the broader medical community’s understanding of the condition they’re suffering from: in the fifth and most recent edition of the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (DSM-V), published in 2013 after a decade of research, ‘gambling disorder’ was for the first time listed alongside disorders related to alcohol, caffeine, hallucinogens, inhalants, opioids, sedatives, hypnotics, stimulants, and tobacco. It is the only non-substance disorder to be listed in the DSM-V.

This was a marked shift from the DSM-IV, which listed what was then called ‘pathological gambling’ alongside compulsive hair-pulling, kleptomania, and pyromania in a section entitled, ‘Impulse Control Disorders Not Elsewhere Classified’. According to Dr O’Neill, the shift is ‘official acknowledgement’ that a gambling addiction is as ‘legitimate and as serious’ as any substance addiction.

In the relevant chapter’s preamble, the authors of the DSM-V explain the reason for the reclassification. There is now, they write, ‘evidence that gambling behaviors activate reward systems similar to those activated by drugs of abuse and produce some behavioral symptoms that appear comparable to those produced by the substance use disorders.’

Understanding the beast that is addiction — especially an addiction to a behaviour where nothing is even drunk, smoked, snorted, or shot up — is both difficult and uncomfortable. Difficult because it demands a trip into the vast, imposing terrains of human psychology and neuroscience; uncomfortable because it shines a spotlight on the innate vulnerabilities all humans share. It’s really no surprise, then, that there’s so strong a tendency to point the finger at the victim. Doing so is not only easy; it is, as Malcolm Battersby, Professor of Psychiatry at Flinders University and the director of the Flinders Centre for Gambling Research, explains, also a ‘natural defence mechanism’.

‘Even if someone has an accident or gets burgled, there’s almost an immediate subconscious reaction of people blaming the victim, saying it is the person’s fault, even though logically it’s not true,’ Battersby tells me. ‘It’s a way of protecting ourselves from anxiety. It makes us think, That wouldn’t have happened to me. I would have done something different.’

Classical and operant conditioning are the first points of call in understanding how someone can become addicted to gambling on a poker machine.

Classical conditioning was first described by Russian physiologist Ivan Pavlov in the early twentieth century. In a series of famous experiments, Pavlov discovered that if a dog was fed at the same time a bell was sounded, it would automatically associate the stimulus — the bell — with the reward — the food — and would eventually be aroused by and start to salivate at the mere sound of the bell, even if the food wasn’t delivered. Once this happened, the animal was said to have been classically conditioned. Pavlov demonstrated that this conditioning was not reliant on any particular stimulus: it occurred regardless of whether the food was paired with a bell, light, or metronome.

Operant conditioning was first described by renowned American psychologist B.F. Skinner. Skinner was interested in whether the manner in which a reward is delivered affects the speed at which an animal learns and unlearns a particular behaviour. He tested this by having pigeons peck a button to receive a food pellet. In some experiments the pellet was issued every time the pigeon pecked the button — known as continuous reinforcement — while in others the pellet was issued only sometimes and randomly — known as variable-ratio reinforcement.

Skinner found that regular, predictable rewards caused the pigeon to peck the button only when hungry. When the pellets were no longer delivered, the pigeon quickly stopped pecking the button entirely. The opposite, however, was true when the pigeon was exposed to irregular and unpredictable rewards. It would peck the button continuously as if in a frenzy because it was forever anticipating that the next peck might trigger the delivery of food. It would also continue pecking long after the food source had been shut off.

Gambling experts have long reasoned that gamblers, and especially poker machine players, undergo both types of conditioning. Professor Mike Dixon and Professor Kevin Harrigan from the University of Waterloo’s Problem Gambling Research Team wrote in a 2009 article that ‘slot machine play involves both operant and classical conditioning’. The Productivity Commission said the same in 2010. Poker machines, it said, ‘also involve potential conditioning effects … which, together with the capacity for rapid repetition, can encourage sustained gambling’.

In terms of operant conditioning, the random, intermittent nature of wins on a poker machine means a gambler is conditioned in exactly the same way as Skinner’s pigeons; their behaviour is quickly learned and very difficult to extinguish. As Skinner himself said when commenting on whether operant conditioning could be observed in humans, ‘There is a good example of how you can move from the pigeon to the human case because one of the schedules that is very effective with rats or pigeons is what we call the variable ratio schedule, and that is at the heart of all gambling devices [poker machines] and has the same effect.’

Because the random wins on a poker machine trigger visual and audible stimuli, gamblers are also classically conditioned: they autonomically associate the stimuli with rewarding outcomes. Professor Dixon and Associate Professor Harrigan explain that this means that ‘just seeing the machine will begin to trigger the (rewarding) arousal response. Thus, a seasoned gambler approaching a slot machine would be expected to show states of arousal before play has even commenced.’

Both classical and operant conditioning are highly effective at reinforcing behaviour that is difficult to extinguish; working alongside one another, as they do in poker machines, only amplifies this effectiveness. But the two concepts offer only a partial explanation about how and why a person becomes addicted to playing poker machines. A fuller understanding is gained by delving into the neuroscience of gambling.

In The Biology of Desire: why addiction is not a disease, neuroscientist and professor of developmental psychology Dr Marc Lewis explains how human brains, like all animal brains, are biologically programmed to pursue reward. This guarantees survival: it is why we relentlessly pursue food, water, and sex. ‘Brains just do what hundreds of millions of years of evolution have determined to be useful, and that includes identifying things that taste good or feel good to us,’ Lewis writes. ‘The brain distinguishes those things from everything else — the background music of the humdrum world — and propels us to go after them.’

Enhancing our ability to pursue reward, Lewis explains, is our brain’s malleability and formation of habits. This ensures that the best rewards can be repeatedly attained with the least amount of effort. ‘New neuronal pathways, and corresponding patterns of thought and behaviour, start off tentative and fluctuating,’ Lewis writes. ‘But after they’ve been activated repeatedly, fledging pathways get more entrenched, more concretized, and eventually carved in stone, or at least in flesh.’

An area of the brain called the limbic system — colloquially known as the ‘reward system’ — is the central actor in this ancient drama. It consists of two major and related components: the nucleus accumbens and the midbrain. The nucleus accumbens is the engine of all goal-directed behaviour. It processes rewards, determining their degree of importance, and adapts to produce the feelings of attraction, desire, anticipation, and craving that are necessary to motivate us to pursue them in the future. The midbrain is where the fuel that powers the nucleus accumbens — the feel-good neurotransmitter, dopamine — is produced and pumped out from whenever a reward is encountered. How strongly the nucleus accumbens fires, how important it considers a reward to be, and how quickly it adapts depends on how much dopamine is pumped out. As Lewis writes, ‘the more dopamine, the more the accumbens is activated, and the more we experience craving’.

What all of this means, says Lewis, is that animals — including humans — are naturally vulnerable to becoming addicted to experiences or substances that are highly rewarding but non-essential — and, in some cases, harmful. It also means that anything which potently stimulates the brain’s reward system can be considered addictive, as is the case for almost all drugs of abuse. ‘Alcohol and heroin would certainly be less addictive (and a lot cheaper),’ writes Lewis, ‘if they led to experiences that are boring’ — boring, that is, from the perspective of the brain.

The pioneering research of Wolfram Schultz, professor of neuroscience at the University of Cambridge, into dopamine in the 1980s and 1990s provides crucial insights into how gambling is able to potently stimulate the human reward system, even though it requires no substance to be ingested. Using a primate as a test subject, Schultz and colleagues administered a squirt of fruit juice following a visual or audible cue, and measured activity in the animal’s dopamine neurons. Before the animal had learned that the cue predicted the reward, the dopamine neurons only responded when the fruit juice was delivered. Over time, however, the response occurred not when the juice was ingested, but when the cue was triggered.

But, curiously, whenever the cue was changed or altered — even if, for example, the time between the cue and the delivery of the reward was either shortened or lengthened — the response to the juice suddenly returned, and the one to the cue faded. After consistently observing this, Professor Schultz realised that the shift was to do with the prediction of the reward.

‘Dopamine neurons fire particularly strongly to unpredicted rewards,’ Schultz tells me. This is, he adds, a survival tool that has developed over millions of years of evolution. ‘It’s a teaching signal. If you do something and suddenly you get a reward that you didn’t expect before — you’ll come back to it, you’ll do it again.’ Likewise, Schultz says the disappearance of the response to the reward and the shift to the stimuli isn’t an evolutionary accident, but a way to anticipate and motivate us to pursue known rewards.

Now think of Doug’s dopamine neurons as he sits down to a More Chilli machine, feeds it a fifty, and hits ‘Spin’. The reels start galloping. As they land and the symbols appear, his neurons desperately try to predict the pattern, exactly as they have evolved to do. Two chillies on the first two reels on this spin means a feature on the fifth spin from now. One money bag on the third reel on this spin means two chillies on the last two reels on the next spin, which means the feature on the third spin after that. But they fail every time, for the simple reason that there is no pattern — every outcome of every spin is always completely random. Every win is therefore always unpredicted, causing a much larger spurt of dopamine in his brain than if the win was predicted.

Of course, this will occur in any game of chance. Roulette, dice, or a lottery — they all award unpredicted wins that will produce an elevated response in the brain’s reward system. But according to Dr Luke Clark, a neuroscientist specialising in gambling addiction, and the inaugural director of the Centre for Gambling Research at the University of British Columbia, poker machines possess certain structural characteristics that make them more arousing and rewarding — and thus more addictive — than other forms of gambling. ‘Wherever you look around the world, slot machines are among the more harmful forms,’ he says. ‘And at the other end of the spectrum you generally have forms like lottery. I think it is the case that slot machines are among the most harmful and more addictive forms.’

Speed is the first structural characteristic of poker machines that Dr Clark identifies. There are two aspects to this, Dr Clark says: first, the length of the anticipation period from bet to outcome — that is, the reel spin or the length of the horserace — and, second, the delay from the outcome to being able to place the next bet. ‘Pokies are fast — short durations — and continuous; lotteries have much longer delays and are not continuous,’ Dr Clark says.

There are no brain-imaging studies looking at the exact effects of these speed variables on the brain’s reward system. This is, Dr Clark says, for a very simple reason. ‘This kind of research is practically challenging because the response we measure with [functional magnetic resonance imaging] is a sluggish response occurring over a five-second window, so it is not really possible to image brain responses to a realistic poker machine, where bets tend to occur every three seconds.’ But using Professor Schultz’s findings about unpredicted rewards, Dr Clark says it is reasonable to assume that the very frequent wins a poker machine player experiences, and the absence of any downtime, results in very frequent and sustained stimulation of their reward system which, in turn, accelerates brain change.

Near misses are another ingredient of poker machines that Dr Clark believes contributes to their addictiveness. Skinner hypothesised this in 1953 when he wrote in Science and Human Behaviour, ‘Almost hitting the jackpot increases the probability that the individual will play the machine, although this reinforcement costs the owner of the device nothing.’

While near misses occur in all forms of gambling, the difference with poker machines is that they are deliberately programmed to occur far more frequently than they would by chance alone. A 2009 paper co-authored by Dr Clark reported that they elicit a response in the brain that is comparable to that of an actual win. ‘The recruitment of win-related [brain] regions during near-miss outcomes underlies their ability to promote gambling behaviour,’ Dr Clark and his colleagues wrote.

Other evidence supports this. In 2017, researchers at the University of Waterloo conducted an extensive literature review of studies that have looked into the effect of structural characteristics of poker machines on players. They found ten studies done between 1990 and 2015 which ‘showed that near misses were associated with large skin conductance responses, a traditional indicator of physiological arousal’. The researchers also found that ‘analysis of brain activity showed that near misses produce activity in areas related to reward and uncertainty, reinforcing the mechanism through which they may act on the player’.

Dr Clark tells me that the brain’s automatic response to a near miss is an evolutionary tool exploited by games of chance such as a poker machine:

Near misses in skill-based games are useful information. If you get very close to the bullseye in archery or if in soccer, you get a shot that bounces off the crossbar, you don’t get the reward but it does still tell you, validly, that your skill level is quite high, that you should persist, that you might be close to getting the goal. So it makes sense that the brain should give those events some value and pay attention to them, even though it hasn’t got anything from that single incident.

When you move to a game of pure chance, the brain treats the events in exactly the same way, only in a game of chance that logic completely falls apart. If your lotto numbers tonight are one away from all of the winning numbers, that doesn’t tell you anything about which numbers you should choose next week. It’s completely invalid but our brain approaches it as if it’s skill relevant information.

The LDW phenomenon in modern multi-line poker machines is the final ingredient that Dr Clark argues is cause for concern. He describes them as a ‘quite devious’ feature.

People in the industry describe them very differently. According to Ryan Jacobs, the insider who suffered from a gambling addiction in his youth and who compares poker machines to women, LDWs are ‘very clever’ and ‘a good concept’ — good, he clarifies, for manufacturers and operators. ‘It’s not being deceptive — it’s right there in front of you. Credits, bet, win — it’s right there. All of a sudden, the player is saying, “I bet a dollar, but only lost 60 cents — that’s not so bad, is it?” Is that being deceptive? It’s not hidden from anyone.’

Dr Clark, however, rejects this argument. ‘LDWs seem to directly cause a player to misperceive what’s really going on in the game,’ he says:

The industry may turn around and say, “Well, the credit window is clearly displaying exactly what happened. We aren’t giving any money or taking away any money that you haven’t won. It is telling you the truth.” But that’s a relatively small area of the screen, and we are not really sure how often players direct their attention to it. Regardless, all of the jingles and feedback that is coming from the machine is telling you that you’ve won when you haven’t.

Of most concern is how frequently this happens in multi-line-format games. If all twenty lines are activated on Aristocrat’s Dolphin Treasure machine, for example, LDWs will occur in over one-fifth — 22.75 per cent — of all spin outcomes.

Evidence to date suggests that the jingles and feedback associated with LDWs cause a significant response in a player’s body and brain.

In 2010, Professor Harrigan and Professor Dixon from the University of Waterloo examined the physiological effect of LDWs. They measured participants’ skin-conductance levels while playing an actual poker machine game, correctly predicting that the winning sights and sounds of LDWs would cause players to react as if they were wins: ‘Our results show that gamblers become equivalently aroused following a win or an LDW, but were less aroused following a loss.’ Professor Harrigan and Professor Dixon added that, ‘despite being losses, LDWs engender the reinforcing arousal that is a key factor in the development of problem gambling’.

While not explicitly about LDWs, a more recent study from 2016 by Professor Catherine Winstanley and Michael Barrus from the University of British Columbia provides further evidence that LDWs effect how a player perceives the game and behaves.

Winstanley and Barrus built what they called a ‘rat casino’, and taught 32 rats to gamble in order to win food. The rats were offered four different gambling options that varied in terms of the associated risk and reward. When there were no lights or sounds, the rats rejected the high-risk, high-reward option, and selected instead the option that awarded small rewards but carried less risk. This was the most rational, sensible decision: in the long term, it was the best way to maximise their ‘profits’.

The addition of lights and sounds, however, dramatically altered the rats’ behaviour. Large, high-risk rewards were accompanied by more celebratory feedback that, in turn, saw the rats pursue those rewards more than small, low-risk ones. This behavioural change intensified when the researchers gave the rats a drug that activated dopamine receptors, and weakened when the researchers administered a dopamine inhibitor. Moreover, those rats that gambled without any celebratory feedback were unaffected by the dopamine drugs.

According to Dr Clark, the findings of this study are highly relevant when considering the effect of LDWs. As he explains, ‘While the Winstanley task does not deliver LDWs, if we assume that LDWs are effective because the sensory feedback causes them to look like a true win, then the Winstanley study indicates a likely dopamine basis to the LDW effect.’

Dr Clark says that not one of the listed characteristics of a modern-day poker machine can be singled out as a sole explanation for their elevated addictiveness compared to other forms of gambling. In isolation, he says, it’s likely that each ingredient produces a relatively mild effect in the brain’s reward system; together, however, they produce a potent and recurring effect.

‘You could come up with a list of twenty or thirty features that might contribute to the addictiveness of modern slot machines, and try to work out if it’s one particular ingredient like the speed or the LDWs, or whether it’s a mix,’ he explains. ‘I think about it like a McDonald’s Big Mac burger. If you asked, “What is the particular ingredient of this wonderful hamburger product that makes it so tasty?”, it’s not going to be one ingredient — it’s going to be a combination of many different features that make it an appealing product. It’s the same with slot machines.’

By focusing on the potential addictiveness inherent in poker machines, Dr Clark is not suggesting that other factors aren’t involved in a person developing an addiction. ‘I do strongly believe there are individual risk factors — things like high stress, loneliness, depression, poverty, poor education, impulsivity, or a hypersensitive reward system — for problem gambling that puts some individuals more at risk than others.’ This, Dr Clark adds, is the case with any addiction, no matter what it is to, and is indeed why only a small portion of drug users or gamblers develop a problem. ‘You can always see a lot of recreational users of a drug who don’t seem dependent on it,’ he says. ‘We now think that with any drug of abuse, only around 15 to 20 per cent of users seem to meet the features for addiction.’

Dr O’Neill agrees that individual risk factors must also be considered. She adds, however, that ‘it’s really important to separate the factors that lead someone to gamble, and the factors that maintain it.’ In her view, individual risk factors are largely what lead someone to gamble, while the characteristics of machine design are what sustains their gambling.

Dr Clark says that focusing exclusively either on the individual or on the machine ‘oversimplifies what is a very complicated relationship between causation and correlation’. According to him, there is a pressing need among experts, policy-makers, the media, the industry, and the general public for more of an appreciation of the nuanced aetiology of gambling addiction:

It is the case that there are some contributions here at the level of the individual. But rather than thinking about that as 100 per cent of the story, we should accept that that’s only about 50 per cent of the story. And that’s a very hard thing to convey in public messages. I don’t want to take the media and the public back to, ‘It’s all about the vehicle.’ I just think we should try to take them back to a rational, ‘It’s both the vehicle and the driver.’ But people don’t tend to like that message, because it is so muddy.

Dr Clark believes that there is an equally pressing need for more research into the neurological effects of certain design traits of poker machines. He admits that the research in this field is still in its infancy and that there are gaps in the knowledge base that need to be filled — gaps not about whether the reward system is affected, but exactly how it is affected.

Conducting this research is challenging — both because poker machines are so fast and sophisticated that examining their effects with current scientific technology is difficult, and because designers are constantly adding new features. As Dr Clark explains, ‘We comment all the time in the labs that we feel like we’re trying to reverse-engineer these incredibly sophisticated games that have evolved over the last 20, 30 years, and are still continuing to evolve to this day.’

Professor Catherine Winstanley, who conducted the ‘rat casino’ study, agrees. Speaking to Wired magazine in 2016, she said, ‘I often feel that scientific models are decades behind the casinos.’