Skip to content

Cognitive enhancement, legalising opium, and cognitive biases

Suppose you want to enhance your cognition. A scientist hands you two drugs. Drug X has at least 19 controlled studies on the healthy individual showing it is effective, and while a handful of studies report a slight increase in blood pressure, another dozen conclude it is safe and non-addictive. Drug Y is also effective, but it increases mortality, has addiction potential and withdrawal symptoms. Which one do you choose? Great. Before you reach out for Drug X, the scientist warns you, “I should add, however, that Drug Y has been used by certain primitive communities for centuries, while Drug X has not.” Which one do you choose? Should this information have any bearing on your choice? I don’t think so. You probably conclude that primitive societies do all sort of crazy things and you would be better off with actual, double-blind, controlled studies.

Now what if I told you that, regardless of your interest in cognitive enhancers, you have been choosing Drug Y over and over, day after day, for several years?

A review of the scientific research on modafinil reveals it produces an enhanced performance on tests of digit span, visual pattern recognition memory, spatial planning and stop signal reaction time; lower error rate in a visual spatial task; increased new-language learning;  improved fatigue levels, motivation, reaction time and vigilance; improvement on spatial working memory, planning and decision making at the most difficult levels, as well as visual pattern recognition memory following delay and subjective ratings of enjoyment of task performance; and so on and so on. All in all there are at least 19 randomized, placebo-controlled, double-blind studies on the effects of modafinil in the healthy individual. No severe adverse effects were reported in most of these studies. Those studies that kept track of blood pressure and heartbeat found none to very little elevation of these measures. Other studies in non-healthy patients have found some adverse effects, but have confirmed modafinil’s safety and – so far – lack of addiction potential.

On the other hand, when we look at the scientific evidence concerning caffeine, although its beneficial effects on overall health are also documented, the average adult male’s dosage surpasses the healthy dosage fourfold. At the average ingested dosage, caffeine has detrimental health effects, increasing all-cause mortality, and also possesses addiction potential, with severe withdrawal symptoms such as depression, irritability, pain and narcolepsy. It considerably increases blood-pressure, the biggest preventable risk factor for death. Long-term use of caffeine has detrimental effects on long-term memory and several studies found it has no beneficial effects on cognition when comparing long-term users with non-users.

What’s more, in general modafinil studies are of better quality, relevancy and maybe trustworthiness than studies on caffeine. Most modafinil studies are new, double-blind, controlled, and targeted at the actual used dosage; whereas most caffeine studies are old and targeted at the wrong dosage, although with a bigger sample size. Additionally, given that modafinil is seen as a dangerous prescription drug and coffee as a harmless habit, we would expect modafinil to be subject to much more scrutiny than caffeine. As coffee is a multi-billion dollar business, it would be tricky to claim that studies have a “funding bias” towards modafinil – if anything the bias would be in the other direction.

The fact that most studies that find caffeine consumption to have a beneficial effect overall use a dosage several times lower than the average consumption is routinely ignored. On the rare occasions where people actually engage with the proper scientific evidence, their conclusion is that there is not enough evidence to say with confidence that modafinil is safe, and that caffeine has been around for centuries. The thesis here seems to be that what really determines something to be safe is how long we have used it. So the fact that scientific evidence shows modafinil is safe and effective is rendered moot because it is a new drug and the scientific evidence that caffeine is unsafe is outweighed by our centuries of experience with it. But cigarettes and alcohol are also pretty old, yet they have killed more people than all the 20th century’s genocides put together. Opium has been used for at least 3500 years, possibly dating back to prehistoric times. Meanwhile, coffee has been around for a mere 500 years. I imagine humanity would not have benefited from using the older-is-safer heuristic in this case. Should we have waged war on coffee and commercialized the production of opium?

There is a set of biases at play when making the comparison between modafinil and caffeine. Why do people use the older-is-safer heuristic? Status quo bias is a consistent and unjustified tendency to prefer that something stays the way it has always been, resisting attempts to change to a better alternative. People continue to prefer a known drug with many side effects over a new, safer drug. Why do people seem to ignore the scientific evidence? It’s because we fail to update our beliefs correctly when presented with absolute probabilities (e.g. 10%) such as the ones provided in most scientific studies. This failing impairs our ability to use information from scientific research to adjust our behaviour. It’s easy to comprehend the risks involved with a certain drug if a friend suffered a heart attack from using it. But reading an abstract number showing that the rise in blood pressure – the most important preventable risk factor for death – of caffeine users is much higher than that of modafinil users is too far away from the experience-based way that our brains are accustomed to absorbing information.

Share on

16 Comment on this post

  1. I think a lot of ‘older is better’ comes from a concern that
    a) There may be uncommon negative effects that only get spotted when a substance is being used by a reasonably large portion of the population – 19 lab studies is not that many compared to the huge body of evidence we have from the population-wide use of caffeine.
    b) Some longer-term effects need time, as well as a large body of people using the substance, to turn up. Modafinil’s been around ~40ish years, with no that many (compared to caffeine) people using it, and not that many (compared to caffeine) really long-term studies. We know that its short-term effects on how our brains work is pretty significant (focus, anti-fatigue), and so we know we’re exposing our brains to something reasonably powerful.

    So while I agree with much of what you say, I would add that I think some of peoples’ new drug aversion is due to a not-irrational fear of rare effects and long-term impacts.

    1. I’m not convinced that using a substance for very long gives us any type of privileged information, except in cases of extreme effects (which the short-term studies partially rule out anyway). Our ability to track down those sort of effects outside controlled studies seems extremely limited. If we take the period of time in which we were no longer putting radioactive uranium on women’s facial cosmetics, cocaine on children’s drinks and things of that sort, then caffeine has only a few decades of advantage over modafinil.

  2. Status quo bias is not a consistent and unjustified tendency to prefer that something stays the way it has always been, resisting attempts to change to a better alternative. There are many ways of defining it, but even the wikipedia entry that you cite has a better one: “Status quo bias is a preference for the current state of affairs.” Status quo bias is almost always tested in the literature in the context of uncertainty, and there are indeed good reasons to prefer the status quo when the outcome is uncertain. Stating that it is unjustified in the face of better alternatives is just a straw man argument. I argued this point as politely as I could when I acted as reviewer for you colleagues’ paper on the topic (Cognitive biases can affect moral intuitions about cognitive enhancement), suggesting to them, as I do now to you, that people will rightly devalue your arguments if you present status quo bias improperly.

    1. I’ll agree with your criticism of the tone in the given ‘definition’, but lord the all bold made it hard to read.

    2. It was a very quick definition indeed and a bit imprecise. However, I maintain it is unjustified in the sense of begin irrational, otherwise it would simply be a good heuristic, not a bias. You might contest that most supposed instances of status quo bias are actually a bias, but it does not seem you can claim a bias is perfectly justifiable. The Wikipedia entry also seems clear on that: “Status quo bias should be distinguished from a rational preference for the status quo ante, as when the current state of affairs is objectively superior to the available alternatives, or when imperfect information is a significant problem.” If you take a look at the scientific literature on status quo bias you will see it is always defined as an unjustified or irrational preference. For instance:
      “Individuals who are subject to the SQB[Status quo bias] tend to choose an alternative that they chose previously (i.e., their status quo), even if it is no longer the optimal choice.” (Kempf & Ruenzi, 2006)
      “… the family of phenomena referred to as status quo bias, which we define as an inappropriate (irrational) preference for an option because it preserves the status quo.” (Bostrom & Ord, 2006)
      Even in the original Samuelson & Zeckhauser paper you will read:
      “The main finding is that decision makers exhibit a significant status quo bias. Subjects in our experiments adhered to status quo choices more frequently than would be predicted by the canonical [rational choice] model.”

      It is also not the case status quo bias is almost always tested in the context of a particularly above-average uncertainty. Many of the classical examples do not involve uncertainty, such as the actual case of electric power consumers in California. Evidently, all cognitive bias are studies in the context of decision under uncertainty, so there is some level of uncertainty. But it is not the case this is the key factor explaining the bias.
      In cases in which individuals have a justifiable, rational preference for the status quo, then it is simply not a bias, it is a rational decision. Nonetheless, there seems to be a great deal of empirical evidence that humans often have an unjustifiable, irrational preference for the status quo.
      If one were to omit that status quo bias is an irrational preference for the current state of affairs, it would be for it being obvious – if it is really a bias, surely it is irrational on some relevant level – not because it is an improper definition.

  3. I’ve taken modafinil. I’d say that yes, it does improve the things listed, but only marginally. It doesn’t make you super special, just ever so slightly better. The one thing I did notice though was that I felt more misanthropic. I had less patience for people. In general, it just made me a little more bitter and I acted a bit like a jerk. I could barely hold conversations without feeling contempt for the other person. I also felt, well, weird. It’s hard to describe. All in all, I don’t miss it. It had some pros and some cons. For me, the latter outweighed the former. Others might feel differently. Maybe I’d take it again if it were OTC and cheaper. Overall though, it’s not like people are missing out on some sort of “limitless” style drug though.

    1. I think actually having experience with the effects of the drugs can be helpful. Too many people think Limitless is possible, and then get rather disappointed. And knowing how the drug interacts with ones quirks is quite useful – just because the drug works well on average doesn’t mean it helps you in your tasks.

  4. I don’t believe the current data we have on consumption of caffeine beverages can draw you to the conclusion that they increase mortality. Your premise is invalid.

      1. This study can not prove causation. “In this large cohort, a positive association between coffee consumption and all-cause mortality was observed in men and in men and women younger than 55 years. ” To prove causation you would need a high-quality prospective study which to my knowledge doesn’t exist. Hope that helps.

        1. That’s true, but it still is an evidence for coffee increasing mortality and it is the big chunk of evidence we have. And given they controlled for must major confounders such as physical and cardiorespiratory fitness, I’m a bit curious of what sort of mechanisms would have be going on here so that coffee wasn’t causing the elevated mortality

          1. Maybe personality traits of users who drink a lot of coffee? Maybe those people are more likely to engage in risky behaviors?

      2. Verbatim from that study site “Drinking coffee has become a normal daily routine for more than half of Americans and large numbers of people worldwide. According to the latest National Coffee Drinking Study from the National Coffee Association, approximately 64% of American adults drink coffee each day, and among coffee drinkers, the average coffee consumption in the United States is 3.1 cups per day.1 Nevertheless, coffee has long been suspected to contribute to a variety of chronic health conditions. During the past 4 decades, the association between coffee consumption and chronic health outcomes has been investigated in relation to conditions such as obesity,2, 3, 4, 5, 6 hypertension,7, 8 and coronary heart disease.9, 10 However, studies on coffee consumption in relation to all-cause and cause-specific mortality are limited, and the results are often controversial. Several studies have found a positive association between higher levels of coffee consumption and all-cause and cardiovascular disease (CVD) mortality,11, 12, 13 whereas others have found an inverse association with all-cause mortality in men and women,14, 15, 16 in women only,17, 18 or in men only,19, 20, 21 with some evidence suggesting that there may be a U- or J-shaped relationship between coffee drinking and health outcomes. Still, other researchers suggest that the association may not exist at all.22, 23, 24 The objective of the present study was to investigate the effect of coffee consumption on all-cause and CVD mortality in the Aerobics Center Longitudinal Study (ACLS) cohort, with average follow-up of 16 years and a relatively large sample of men and women.”

  5. I think there are a number of question marks about the evidence for modafinil being an effective enhancer. We can cite individual studies that report an effect, but these are often too small and statistically underpowered. We should also be wary of the fact that null results are not likely to be reported — these kinds of studies are rarely registered in advance. On the other hand, the picture from systematic reviews and meta-analyses is less promising (e.g. Chamberlain et al. 2011; Repantis et al. 2010) . It seems to me that, at best, all we can say about modafinil at the moment – in terms of its effects on healthy, non-sleep deprived individuals operating in laboratory conditions – is that the evidence is mixed.

    Also, it is not entirely clear how addictive or non-addictive modafinil is. No long term studies exist (as far as I’m aware most last 2-12 weeks), while more evaluative studies have tiny samples (9 in the one you cite). The fact that it appears to release substantial amounts of dopamine similar to those released by other psycho-stimulants has worried at least some about its potential for long-term abuse and addiction (e.g. Heinz et al. 2012)

    Finally, unlike caffeine, at the moment it doesn’t seem entirely clear what brain mechanisms modafinil is acting on, making it appear potentially riskier long-term. Altogether, it seems to me that it’s not necessarily that people are ignoring the scientific evidence, but rather that the scientific evidence is not as clear cut as suggested here.

    Chamberlain, S. et al. 2011. “Translational Approaches to Frontostriatal Dysfunction in Attention-Deficit/hyperactivity Disorder Using a Computerized Neuropsychological Battery.” Biological Psychiatry 69 (12): 1192–1203. doi:10.1016/j.biopsych.2010.08.019.

    Heinz, A., et al. 2012. “Cognitive Neuroenhancement: False Assumptions in the Ethical Debate.” Journal of Medical Ethics 38 (6): 372–75. doi:10.1136/medethics-2011-100041.

    Repantis, D. et al. 2010. “Modafinil and Methylphenidate for Neuroenhancement in Healthy Individuals: A Systematic Review.” Pharmacological Research 62 (3): 187–206. doi:10.1016/j.phrs.2010.04.002.

    1. It is not the case the evidence for modafinil’s effectiveness as a cognitive enhancer is mixed at all. There are only two studies I know of which didn’t find an effect (one was just out). There are many studies with over 60 participants (e.g.: Muller, 2012; Turner, 2003) which did find a significant effect and were not statistically underpowered. The bias against null results would be even stronger for caffeine, where many studies are funded by the coffee industry.

      It is also not the case the evidence you cite support your claims. The only meta-analysis you mentioned (Repantis, 2010) actually concludes: “Modafinil on the other hand, was found to improve attention for well-rested individuals, while maintaining wakefulness, memory and executive functions to a significantly higher degree in sleep deprived individuals than did a placebo.” Moreover, this meta-analysis also left out many studies that weren’t published at the time. If you take a look at the references I cite, most of them are post-2010. I would expect that if they had included those studies they would have found an even bigger effect.

      It is worth noticing I’m not comparing modafinil with nothing. Although it might not be entirely clear that modafinil is not addictive, it is clear caffeine is. Therefore, when comparing the two one should prefer the former. There are several long-term studies of modafinil in patients with narcolepsy, none has found addiction potential, or evidence of tolerance even after one year of use (quite unlike caffeine). If memory serves me well there is at lest one 6-months study of modafinil on healthy individuals which didn’t find any major side-effects. There is also little reason to expect it would be non-addictive in the narcoleptic and addictive in the healthy(actually, the opposite is the case for many drugs). Dopamine release alone is not a ground for concluding a substance is addictive, and dopamine’s role on modafinil’s effects might be little. Additionally, there are many people taking modafinil for narcolepsy for dozen of years and no addiction behaviour has been reported (quite unlike caffeine). Evidently, modafinil has risks, but so far the scientific evidence is that is less risky than caffeine.

      I state again that people are in fact ignoring the scientific evidence.

      1. Note that I cited Repantis et al. (2010) alongside another meta-analysis (Chamberlain et al. 2011) to highlight just how mixed the evidence is: while Repantis found modafinil had a moderate effect on attention, Chamberlain found no effect on attention. The Chamberlain analysis highlights just how inconsistent the evidence is, not to mention that on the five measures it tested for which there was sufficient data, it found no evidence overall for significant effects of modafinil among healthy participants.

        Regarding Turner et al. (2003), note that – in contrast to some of the studies you cite — it found modafinil had no effect on spatial working memory, attentional set-shifting task performance, or rapid visual processing. Where it did find an effect, attempts to replicate their positive findings using the same tasks with healthy non-sleep deprived subjects found no effects (Randall 2004; 2005a; 2005b). Note also that the Müller study you cite failed to find any effect on digit span/maintenance, as well as a number of other tasks.

        It’s also worth mentioning another study that found modafinil had no effect on cognitive function, but only acted to increase anxiety (Randall 2003). Finally, the latest study on modafinil, which I believe you mention in passing, found no evidence that it boosts cognition, but that it actually slows down response time (Mohamed & Lewis 2014).

        Taken together, I don’t know how we can still conclude that the evidence for modafinil isn’t mixed. It clearly is. But one point worth conceding is that I suspect if we look at the evidence for some of the most commonly used drugs on the market, the evidence from studies wouldn’t necessarily be much more consistent. Moreover, I acknowledge that your point is merely to contrast it with caffeine.

        But even with that in mind it is worth noting what limited evidence we have: Wesensten et al. (2002) compared modafinil and caffeine directly and concluded that modafinil does not appear to offer advantages over caffeine. A later study comparing modafinil, dextroamphetamine and caffeine found similar effects from modafinil and caffeine (Wesensten et al. 2005). But the study notes that the shorter half-life of caffeine can make it more practical since it can allow users to make use of an opportunity for sleep shortly after taking it, should such an opportunity arise. Not so with modafinil’s longer half-life.

        Ultimately, I nevertheless think the evidence is probably sufficient to make modafinil more widely available and less controlled. In the meantime, I have no doubt that some status quo bias is at work. However, this does not change the fact that the evidence is by no means as unambiguous as you suggest. Acknowledging that can play a part in protecting the broader enhancement literature from (currently spot on) accusations of hype.

Comments are closed.