This chapter:

• introduces neuroscience aims, claims and translations;

• introduces scientific approaches, and discusses belief, doubt and denial;

• highlights the need for critical thinking to both monitor and defend science; and

• outlines the relevance of philosophy and history to science.

I have been a neuroscience researcher for thirty years, principally working on spinal cord circuits following an interest in nervous system plasticity that developed during my PhD. Conventional measures show my work has had little impact: I have not published in any high-impact journals, I have been cited ~2,200 times and I have a H-index of 27 (i.e. twenty-seven of my papers have been cited at least twenty-seven times) – all modest achievements. Removing my contribution would have little impact on my field (I am rarely cited so this situation effectively exists). My significant research grants ended by 2006, as did invitations to give talks at universities and conferences: my career stalled at this time before going into decline.

My lack of prominence is a fair reflection of my contribution. I have published work I have backtracked on (I discussed one class of synapse in three publications, changing my conclusion each time).Footnote ¹ I have made analytical mistakes and used sample sizes that I now know lacked statistical power, and removed what I considered were ‘outliers’ from data, later realising they reflected natural variability.Footnote ² I repeated a figure in two papers from my PhD that showed different things (I think because I wrote them at the same time and got mixed up).Footnote ³ I have made logical and experimental errors, the latter typically reflecting errors in making solutions. The resulting effects showed something was obviously wrong, but did more subtle errors go unnoticed? Some of my work has been replicated, some hasn’t, which may reflect natural variability or differences in methods, or indicate that my data is unreliable.

My grants ended and my research group dissolved soon after I got my position at Cambridge. I stopped applying for funding because it was clear I wouldn’t get any. My papers also started being rejected, so I lowered the profile of the journals I submitted my work to. I essentially stopped trying to satisfy the career goals of a scientist of getting funding to run a research group and trying to publish with maximal impact. Freed from these demands, I have largely worked alone, following my own interests. This has been the most enjoyable part of my career. This is a luxury few have, made possible because Cambridge is still sufficiently free of the business model of science to allow it.

Given this summary, why should you read on? Firstly, in my defence, errors are common and I have seen others make the ones I have made. Secondly, most scientists, including those who achieve prominence, ultimately make only modest contributions: scientific advances typically reflect the accumulated input from many people. Finally, I started my PhD in the 1990s, which now seems like a transitional period in neuroscience: molecular techniques were introduced that changed assumptions and experimental approaches, and neuroscience research changed from work done by individuals or small labs to collaborative ‘big science’ projects (e.g. the Human Brain Project and the US BRAIN initiative) that generate multi-author papers.

But the principal reason for this career summary is because I want to emphasise the need to question rather than simply learn material, and this includes mine. Deference to authority is a powerful barrier to thinking critically, one that my lack of prominence should remove.

I was once told, ‘we know you don’t believe anything’. I think this was meant as an insult, but should scientists have beliefs? Scientists aim to describe what is real, not what we believe (the philosopher Ludwig Wittgenstein dismissed belief from science because it can become separated from reality). The opposite of belief is doubt or denial. There is an important difference between these. The French physiologist Claude Bernard claimed, ‘The doubter is a true man of science: he doubts only himself and his interpretations, but he believes in science.’Footnote ⁴ Doubt was also emphasised by the nineteenth-century German physiologist Emil du Bois-Reymond, who in an 1872 speech titled ‘On the Limits of Science’ listed several ‘riddles’ that reflected things science either didn’t know or he doubted ever would.Footnote ⁵ Doubt is also inherent in the philosopher Karl Popper’s claim that scientists should try to falsify rather than verify their hypotheses, something generally presented as the scientific method. Doubt can promote discussion that detects and correct errors, strengthens interpretations or suggests new analyses that can lead to progress; denial refuses to discuss or acknowledge alternative views, closes debate and restrains a field.

As a neuroscience researcher, I have seen assumptions accepted as understanding, and adherence to certain beliefs but denial of others (this is, of course, subjective). As a lecturer, it seems that teaching has become more about delivering material for students to learn and assessing them by how well they remember it, rather than fostering an approach that considers the material. This is also a subjective view, but one supported by student and lecturer feedback. Learning ‘facts’ is less important now than at any time in history, given that for many people this information is available online. Science was never about learning details (this is more relevant to medicine and engineering, where appropriate responses to specific conditions need to be learnt) but about knowing what a field knows, its open questions and how to address them. In claiming to not know the speed of sound, Einstein said he didn’t ‘carry such information in my mind since it is readily available in books … The value of a college education is not the learning of many facts but the training of the mind to think.’Footnote ⁶

Neuroscience is a diverse subject that covers nervous system development (developmental neurobiology); the molecular and cellular mechanisms of neuronal and synaptic function and plasticity (molecular neuroscience and neurophysiology); and psychology and behaviour (systems, behavioural and cognitive neuroscience). These fields use diverse techniques, including molecular genetic approaches; anatomical analyses, using various forms of light and electron microscopy; electrophysiological recordings of single neurons, using whole cell or intracellular recordings or of nervous system regions, fibre tracts or nerves, using extracellular recordings, including electroencephalography (EEG); calcium or voltage imaging of single cells or imaging of brain regions (e.g. functional magnetic resonance imaging, fMRI); and mathematical and computational approaches (computational neuroscience).

I won’t go into the detail of these fields or techniques; these can be found in any neuroscience textbook or online. I will instead use various examples to illustrate the need to think critically, to evaluate experimental approaches and claims to explanation and understanding. I will do this largely from the perspective of neuronal circuits, groups of neurons with specific molecular and cellular properties that influence sensory, motor and cognitive functions. These circuits occupy a central position in neuroscience: if a developmental or molecular neuroscientist wanted to know how the aspects they study relate to behaviour, they would move up to the circuit level; and if a psychologist wanted to know the neural mechanisms underlying the behaviours or cognitive effects they studied, they would move down to the circuit level. Even if neuronal circuits are of no interest, they relate to the general issue of how parts relate to wholes: this is ‘mereology’, an aspect that is relevant from physics to sociology.Footnote ⁷

Knowing how molecular and cellular properties link through circuits to normal and abnormal nervous system functions and behaviours is frequently claimed to be the biggest problem facing neuroscience.Footnote ⁸ Wikipedia has a ‘List of Unsolved Problems in Neuroscience’ that largely reflect circuit aspects, including the neural basis of consciousness, memory, sensory processing and movement.Footnote ⁹ The Nobel Prize-winning neuroscientist Eric Kandel said that linking molecular and cellular components to behaviours was among ‘the most difficult problems confronting all of biology – in fact, all of science’.Footnote ¹⁰ While you won’t struggle to find statements like this, you also won’t struggle to find people claiming that these links have been made (including Kandel; see Chapter 7).

With claims to knowledge and understanding come claims to translation. This is a significant consideration for any science, and obviously so for neuroscience, given the impact of erroneous interventions in the brain. Knowledge of the central nervous system allows neurosurgeons to effectively treat tumours and aneurysms and address the effects of stroke and head or spinal cord injury, while neurologists can diagnose conditions and address the symptoms of neurological disorders or delay their onset by targeting biological aspects known to be associated with specific conditions. Biological markers are lacking for psychiatric conditions and diagnoses are instead made from cognition or behaviour. While this doesn’t identify specific biological targets, psychopharmacological interventions target neurotransmitter systems. Despite millions of people using these drugs in increasing numbers,Footnote ¹¹ their mode of action and efficacy with chronic use is still uncertain.Footnote ¹²

The history of physical interventions in psychiatry is poor. These included restraint, sedation, cold water immersion, injection of malaria-infected blood, drug or electrically evoked seizures, insulin-induced coma and removal of teeth or parts of the digestive tract. The best known ‘shock’ approach is lobotomy, in which connections between the frontal lobes and the rest of the brain were severed. This was performed on over 100,000 people in the middle of the twentieth century to ‘treat’ various disorders but left most people severely psychologically or physically incapacitated. A notable case is John F. Kennedy’s sister Rosemary, who was lobotomised following promises it would cure her ‘disruptive’ behaviour. It is uncertain whether this reflected mental retardation following a difficult birth or because she enjoyed an active social life that her father felt could affect his family’s political ambitions. Whatever the reason, the lobotomy left her severely incapacitated and she spent the remaining fifty years of her life in a care home.Footnote ¹³ Lobotomy was used extensively in the West and its originator won the Nobel Prize, but it was banned by the Soviet Union as ‘contrary to the principles of humanity’.Footnote ¹⁴

Translation goes beyond clinical conditions. Parents can be exposed to expensively packaged pop psychology claims that they can enhance their children’s cognitive development. In the 1990s the ‘Mozart effect’ claimed that listening to Mozart improved cognitive performance: it turns out that listening to music generally can do this, probably by increasing arousal. Similar claims are now made about ‘mirror neurons’. These are neurons that are active when a behaviour is being performed or when it is observed being performed by others. These neurons are claimed to have shaped human civilisation, and parents are promised that by targeting them they can ‘tune’ their children’s social and cognitive development, claims that are based on very limited insight into mirror neuron function.Footnote ¹⁵

Playing Mozart CDs or ‘optimising’ mirror neuron function might not have any beneficial cognitive effects in children (although the associated parental attention could), but they won’t have any negative effects either. But a current focus of cognitive enhancement uses drugs to target neurotransmitter systems in the brain that are claimed to underlie specific functions. Any pharmacological intervention in the brain comes with risks that should require we understand the transmitter systems being targeted and their roles, rather than hoping that the intervention will be safe and beneficial. A 2005 report by the UK Office of Science and Innovation ‘based on scientific knowledge’ predicted that by 2017 cognitive enhancers may be ‘part of the knowledge professional’s toolkit’.Footnote ¹⁶ This failed prediction highlights the limited science behind the claims: one drug, donepezil, which reduces the breakdown of acetylcholine, was said to ‘produce more acetylcholine [levels may be higher, but more hasn’t been produced], allowing the brain to work at a higher level of efficiency’;Footnote ¹⁷ while transcranial current stimulation was claimed to ‘make areas of the brain work even harder’.Footnote ¹⁸ What does ‘a higher level of efficiency’ or the brain working ‘harder’ mean?

Negative translations don’t only result from intervening in the brain but also from social interventions. An example is secondary school initiatives to address adolescent mental health problems. While these initiatives are potentially beneficial, assumptions that they will benefit all is negated by evidence that they can increase distress or lead to people thinking they have the symptoms of the conditions in question.Footnote ¹⁹ Another example is the UK’s eleven-plus exam used from 1944 to the late 1970s (and still used in some areas today), which was based on seemingly fraudulent data from the psychologist Cyril Burt.Footnote ²⁰ Burt claimed that intelligence was largely inherited and that educational achievement could be predicted at the age of eleven. Those who passed the eleven-plus exam had an academic grammar school education, and those who failed went into the secondary modern stream, with less funding per student and a limited academic focus (basic academic subjects in the morning, and afternoons of metalwork, woodwork, gardening and sport). The eleven-plus significantly influenced the career prospects of several generations of schoolchildren. While Burt is seldom mentioned now, assumptions of inherited ability are still common. In a 2007 interview, the co-discoverer of the structure of DNA, James Watson, ignored obvious social factors to claim, ‘There’s a difference on the average between blacks and whites on IQ tests. I would say the difference is genetic.’ Watson also claimed the genes responsible for intellect would be found within ten years. They weren’t: you will see that ‘in ten years’ is a neuroscience mantra.Footnote ²¹

Translations are also sought between neuroscience and technology, in particular robotics and artificial intelligence (AI). The benefits of AI are already clear from its use in smart devices, language translation, navigation tools, content creation and medical diagnoses. AI systems can identify patterns in data, hold conversations, drive autonomous vehicles and beat the best human performers at chess and other games. But even under the relatively controlled conditions of games with set rules and training using carefully controlled data sets, AI systems are fragile and can easily make mistakes, including in medical diagnoses,Footnote ²² with ‘model collapse’ seeing them generating nonsense as errors are amplified on successive iterations.Footnote ²³ Despite their obvious utility in specific domains, AI systems are far from able to replicate the general abilities of a human or even an insect brain. This is expressed in Moravec’s paradox, which holds that artificial systems outperform us on aspects we find difficult (e.g. storing, retrieving and processing large amounts of data) but struggle with things we take for granted (e.g. movement).

Investment in neurotechnology has been increasing rapidly since 2010. Yuste et al. claimed

What should we make of these transhumanist claims? When billionaires lacking scientific insight promise superhuman cognition and telepathy, we can roll our eyes, but claims such as these come from respected scientists and philosophers. Despite the authors saying it may take decades, doesn’t this sound like science fiction? We have been able to record brain activity and use it to control artificial devices (e.g. computers or robotic systems and prosthetic limbs) or relay the signals to other parts of the central nervous system (e.g. across a spinal cord injury site) for over twenty years. While these are impressive achievements with obvious potential benefits,Footnote ²⁵ the performance of these systems in controlling even basic movements is still limited. A table tennis playing robot recently developed by Google DeepMind was trained using computer simulations and in actual games using cameras that track the ball’s position and an opponent’s movement (using LEDs on the opponent’s bat). Its performance was claimed to be ‘mind-blowing’, but the robot couldn’t serve: if it won games, it was because they were adapted to what the robot could do.Footnote ²⁶ Even if robotic movements do eventually match human performance, this wouldn’t support the claims of AI-enhanced human mental abilities set out in the earlier quotation.

AI claims have also been considered in ethical discussions of how to protect people’s ‘neural data’ from being used by governments and corporations.Footnote ²⁷ While this won’t necessarily stop people trying, it isn’t clear what this data is or how it could be used. Just as a genome sequence can tell us little unless we know where the genes are and how they are regulated, reading people’s ‘neural data’ requires we know what the activity means. This is complicated by long-standing uncertainty over the nature of the neural code.Footnote ²⁸ Current ‘mind-reading’ approaches see subjects spending many hours in an fMRI scanner to allow imaged activity (i.e. blood flow to different brain regions) to be correlated with certain words and phrases. But even under these constrained conditions, accuracy is limited. A recent study recorded electrical activity directly from the prefrontal cortex in patients undergoing surgery for epilepsy: while specific classes of words (e.g. animals or foods) could be correlated with certain patterns of activity, it still wasn’t possible to decode what people were hearing.Footnote ²⁹

The societal impact of any new technology, especially if it is autonomous, should be considered.Footnote ³⁰ We already know of negative misinformation issues arising from AI-generated text, images and videos, and there are concerns that billions of people are being exposed to technology controlled by a handful of people. A more dramatic claim is that autonomous AI systems will enter runaway cycles of self-improvement that will lead to a ‘technological singularity’ that presents a major existential threat to humanity. The computer scientist and philosopher Hubert Dreyfus called these claims ‘crazy’.Footnote ³¹

AI can be inspired by the human brain to varying degrees, from ignoring brain processes to emulating (mimicking) them. Existential threats of brain-derived AI seem exaggerated given our current knowledge of the brain and that adaptive or intelligent behaviour doesn’t simply reflect the brain but its embodiment in a body in an environment of other embodied brains. But the threat of the singularity is receiving serious academic and political attention: for example, Stephen Hawking claimed that ‘full artificial intelligence could spell the end of the human race’.Footnote ³² This threat is being promoted by the AI industry itself,Footnote ³³ with some claiming that it is a smokescreen to distract attention from the need to deal with the current issues of misinformation and monopolistic control, or an attempt to persuade governments to enter and invest public money in a corporate ‘AI-race’ by promoting the view that the most pressing issues facing society are technological, or that social and political issues are better addressed by AI technology and the free market – that issues affecting humanity reflect engineering or algorithmic problems. In 2023, the UK government made addressing the technological singularity and AI a focus,Footnote ³⁴ maybe to distract from issues they should have been dealing with. There are various predictions of when the technological singularity will occur: some dates have already passed, some are imminent (2030–2050), and some people follow Dreyfuss and say it is unlikely to ever happen.Footnote ³⁵

Claims to understanding and translation highlight the need for critical thinking and debate, namely, to: ensure that the claims are not obviously wrong, exaggerated or misleading; lead to erroneous investment of time and money; or promote premature interventions. Wealth may give power, but we have enough examples to know that it doesn’t necessarily equate to insight or intellect. We can let billionaires and investors spend their money on AI and other technology research that will undoubtedly lead to many profitable gadgets and apps, but we should hold AI claims to account to ensure they don’t influence governments to (mis)direct policies, research or public money on hype, or lead to interventions that lack the necessary scientific basis. This will make the interventions a lottery, and any lottery has far more losers than winners.

Critical thinking isn’t only needed to evaluate scientific and technological claims and translations but also to defend science against denial and non-scientific views. Public denial of science seems to reflect a lack of trust in science, not a lack of scientific knowledge, as often assumed.Footnote ³⁶ The latter assumption can see media presentations of science being dumbed down and superficial, which can lead to misunderstandings or present a false balance where dichotomous views are given equal attention, despite the evidence falling strongly on one side (e.g. climate change or discussion of evolution balanced by religious beliefs).

Political and corporate science denial can reflect economic or ideological factors that span the political spectrum. A 2023 survey in the United States showed that only 49 per cent of Republicans and 69 per cent of Democrats felt that science positively benefited society.Footnote ³⁷ Robert Kennedy Jr, traditionally a Democrat, is a vaccine denialist, and left-leaning academics have been accused of arguing against scientific realism (see Chapter 10). The British Medical Journal reported links between UK Conservative politicians and the Institute of Economic Affairs, a right-wing think tank that receives corporate funding to campaign against public health initiatives (alcohol controls, sugar taxes, tobacco advertising, etc.).Footnote ³⁸ Finally, a 2008 survey showed that over 90 per cent of books published in the United States that promoted disinformation about environmental issues had links to conservative think tanks, and voting against environmental issues by members of the US Congress was proportional to the donations they received from the fossil fuel industry, an industry whose internal documents have acknowledged the reality of anthropogenic climate change for decades while publicly denying it.Footnote ³⁹

Whilst not necessarily nefarious (i.e. intent may be genuine), corporations can use various strategies to promote their interests. This includes financially supporting members of editorial boards of journals to facilitate publication of their research and financing university positions or infrastructure to claim academic support. More nefarious corporate practices are ignoring negative or contradictory results and biasing experiments towards certain conclusions (food industry funded research tends to generate results more favourable to the industry than non-industry funded research),Footnote ⁴⁰ and funding research intended to promote doubt about evidence critical to them or their products. The obvious example is the funding of research into the genetics of lung cancer by the tobacco industry to detract from known links between smoking and cancer.Footnote ⁴¹ Corporations and other groups can also lobby governments for legislation favourable to them.

During the past decade there were over 500 attempts to restrict or misuse scientific research in the United States. Legal challenges have also tried to restrict the teaching of evolution or forced the parallel teaching of creationism and intelligent design by claiming science is just another belief. Of greater concern is the harassment and attempts to discredit scientists who negate corporate, political or religious interests.Footnote ⁴² US government scientists who conflict with ideological views have been demoted,Footnote ⁴³ and Richard Nixon’s claim that professors were the ‘enemy’ was repeated by the Republican vice president JD Vance. Academics have been subject to organised online abuse, lawsuits and media attacks by various groups who want to limit discussion that conflicts with their views. Stanford University ended a collaboration with the University of Washington to document misinformation in the 2024 US presidential election because of the time and cost involved in dealing with these challenges. Republican politicians celebrated this as a victory for ‘free speech’.Footnote ⁴⁴

These campaigns seem to be succeeding in promoting distrust of science and by making false information part of mainstream discussion. In an age of increasing political division, corporate power, and social media and AI-generated misinformation it is more necessary than ever to highlight the illogic of ‘alternative truths’ and misinformation applied to scientific and other evidence. This danger has always existed when information is provided by people with vested interests (e.g. corporate lobbying and advertising, media owners and populist politicians). Scientists shouldn’t flinch from addressing and debating these issues, but debate on both sides has to adhere to standards of evidence and logic, not alternative truths. The best way of addressing misinformation isn’t to police material at source – political and corporate interests are unlikely to respond to rational critique and standards of logic – but instead to foster a general critical attitude, a ‘psychological inoculation’ that allows people to evaluate and negate erroneous claims and views.Footnote ⁴⁵ This approach was promoted in the mid nineteenth century by the opening of UK public libraries paid for by voluntary public subscription, a ‘triumph of unselfishness over the lust of money’ that was opposed by conservative politicians who claimed public education would agitate the masses, and by UK workers’ educational movements that challenged and ultimately changed unfair social and employment conditions.Footnote ⁴⁶

Scientists can help address the lack of trust in science by avoiding hyperbole and premature claims to understanding and translation, as both can weaken trust when the claims fail. Both climate change and Covid pandemic denials refer to model revisions as evidence that the science is unreliable or untrustworthy. Scientific debate can also be misconstrued: an example is reference to scientific debate over phyletic gradualism or more rapid punctuated equilibrium by creationists in their attempts to dismiss the theory of evolution, a claim that fails to appreciate that the debate is about how evolution occurs not if it occurs. These erroneous critiques are easier to negate if people understand that science doesn’t give absolutes just the best current explanation or prediction from the available data, and that these can differ in models that focus on different aspects or change as new data is acquired. This doesn’t make scientific statements arbitrary or unreliable. We wouldn’t claim a doctor who revised their diagnosis after a blood test was untrustworthy but would see this as appropriate and rational caution, and of course people accept (and seek) frequent updates to technology without claiming it must be unreliable.

Karl Popper, whose approach is often presented as the scientific method, said that scientific knowledge is provisional not factual (if factual means necessarily and infallibly true for all time), but by using rational critique we can reject false hypotheses to retain ‘truth-like’ statements: those with the best explanatory or predictive power.Footnote ⁴⁷ Terms like ‘true’ and ‘fact’ are avoided because they suggest a finality that can promote dogma and limit further work, whereas a default of scientific caution leaves open the utility of further discussion and work that can ensure claims are accurate. This caution applies even when scientific hypotheses have practical applications. The erroneous Earth-centred view of the solar system was used successfully in astronomical charts and in navigation for 1,500 years, and the eighteenth-century theory that claimed combustion reflected the release of a substance, phlogiston, rather than combination with oxygen had successful practical applications in the production of sulphuric acid and the isolation of nitrogen and carbon dioxide: despite their practical utility, both theories were wrong. Celebrated examples like these don’t mean that we should distrust science or expect every scientific hypothesis or theory to eventually be negated: for Popper, truth-like status is earned through logical consideration of the accumulated evidence.

There may be field-dependent differences in how readily we accept aspects as truth-like; for example, observational fields like astronomy or anatomy compared to explanatory areas like physiology or psychology. But we should be cautious of claiming factual or truth-like status even in the former areas because long-standing views can still be revised or negated. For example, there are currently eight planets in our solar system, but between 1930 when Pluto was discovered and 2006 when it was downgraded to a dwarf planet there were nine; seemingly established aspects like the definitions of species and the organisation of the periodic table are also being re-considered.Footnote ⁴⁸ A neuroscience example is the textbook view of the autonomic nervous system that controls functions like blood pressure and temperature. Medical students learn the ‘fact’ that this has a cranial and a sacral (lower) spinal cord parasympathetic division separated by a thoracic spinal cord sympathetic division, the two divisions having characteristic anatomical features and neurotransmitter systems. But recent molecular genetic analyses have suggested that the sacral region may be sympathetic, not parasympathetic.Footnote ⁴⁹ Considering our views and revising them, if necessary, advances knowledge and improves translations that are based on it.

To counter lack of trust in science we should present the evidence and outline the reasons why we hold certain views (this is often done), while also emphasising any uncertainties or contradictions (this is less common). While Claude Bernard said that doubt was a part of science, views of science and scientists can at times be over-confident. Consider this description by the neuroscientists Roger Nicoll and Robert Malenka:

These very positive views of science and scientists were contrasted by the astronomer Carl Sagan who wrote, ‘In science it often happens that scientists say, “You know that’s a really good argument; my position is mistaken”, and then they would actually change their minds and you never hear that old view from them again … It doesn’t happen as often as it should, because scientists are human and change is sometimes painful.’Footnote ⁵² The Nobel Prize winning immunologist Peter Medawar also disagreed with Nicoll and Malenka’s claim that scientists rationally follow a ‘clearly enunciated’ scientific method. He said ‘Ask a scientist what he conceives the scientific method to be and he will adopt an expression that is at once solemn and shifty-eyed: solemn because he feels he ought to declare an opinion; shifty-eyed because he is wondering how to conceal the fact that he has no opinion to declare!’Footnote ⁵³ Although Nicoll claimed in the quote above that Popper’s scientific method is ‘the very essence of scientific inquiry’, the autobiographical account of his scientific career suggests he didn’t follow this approach. It instead refers to ‘confirming’, ‘finding’, ‘showing’ and his ‘amazement’ at certain results and his disappointment when predictions failed.Footnote ⁵⁴ Note that none of this negates Nicoll’s scientific contributions: as you will see in Chapter 2, although contradicting his and Malenka’s view of scientists, Nicoll’s autobiographical account is arguably the norm for science.

Sagan and Medawar were not critics but staunch supporters of science who recognised that as a human activity science can be affected by human faults and foibles. In contrast to Nicoll and Malenka’s description of a scientist, Medawar said that scientists are diverse in their approaches and attitudes, and include classifiers and compulsive tidier-uppers, detectives and explorers, artists and artisans, poet-scientists and philosopher-scientists, and a few mystics.Footnote ⁵⁵ He added there were ‘a few crooks’ whose behaviour he said removes everything that makes a scientific career worthwhile, namely the sense of making a genuine contribution.Footnote ⁵⁶ Medawar was referring to people who falsified data, but this would also seem to apply to scientists who exaggerate their claims or close debate to protect them. This seems like not finishing a novel because you have already decided on the ending.

I have personally known highly methodical scientists who plan and follow a series of experiments, essentially completing the figures for a paper as they go, while others take a more haphazard approach, doing things as they occur and pulling the various strands together (this is my approach, one I regret when it comes to data analysis). Some scientists are very organised and precise, some less so. While the latter seems inconsistent with the default view of scientific precision, the Nobel Prize winning biophysicist Max Delbrück claimed a scientific ‘principle of limited sloppiness’: ‘sloppy’ experiments with many uncontrolled variables allow little to be understood, while tightly controlled experiments leave little opportunity for novel discoveries; but experiments that are a little sloppy can lead to unexpected findings and new insight.Footnote ⁵⁷ Alexander Fleming seems the epitome of Delbrück’s principle: unlike his colleagues who cleaned their lab benches daily, Fleming left his bacterial culture plates on his bench for weeks, but he would always check them for anything interesting before clearing them away, an approach that led to his discovery of penicillin.Footnote ⁵⁸

This diversity of scientists and scientific approaches adds to the need to think critically to evaluate data obtained by people with different views and using different approaches. But despite the various reasons highlighted for why critical thinking is needed, it can be difficult to foster. This difficulty can begin at the undergraduate level where students may be reluctant to critique material because they assume that learning what a lecturer says is the optimal exam strategy, while PhD students or postdoctoral researchers will start by learning the basis of their field and the techniques they need to use rather than by considering its views or approaches. In all cases this can lead to dogma being accepted that can constrain future views and work.

My first exposure to critical thinking came from an inspirational lecturer in my final year as an undergraduate. In each lecture he presented evidence for a viewpoint and then highlighted its assumptions, weaknesses and alternatives. Despite this it wasn’t until I was a PhD student that I started to read papers critically. I had to present a paper at a lab journal club but only got to the first figure before my PhD supervisor took over the presentation, pointing out flaws in that figure and the rest of the paper that I hadn’t, and wouldn’t have noticed. This lesson had the intended effect of making me read papers critically. Even then it was a few years later when I was doing postdoctoral work that I really started to think critically. Like many others I had accepted the claim that the neuronal circuit I was working on (the lamprey locomotor circuit) was fully characterised, but the experimental data needed to support this claim is either insufficient or missing (the latter is demonstrable from the literature, the former open to debate, but not debated; see Chapter 7). This is when I started to doubt the textbook view of science: why are claims made that can’t be supported, and more importantly why do others accept the claims when in some cases they privately expressed doubts or ridiculed them? Getting to this point had taken me a decade after I started university. Someone can finish a PhD equipped with various experimental and analytical skills but still lack the ability to think critically; or at least I did.

I am aware that claiming students should question the views of more senior scientists is easier said than done. There are scientists who will accept critique and discuss their claims and ideas, but the diversity of scientists means there are some that won’t. A student or postdoctoral researcher may hesitate to ask someone senior or prominent to defend their views, either because they don’t want to look foolish by not seeing what everybody else sees as obvious, or because they don’t want a prominent person in their field to have a negative opinion of them (this is discussed in Chapter 10). But a prominent scientist who is confident of their claims should be willing to explain why they are confident rather than expecting deference to their authority. If they aren’t then there is little that can be done. But something useful will still have been learnt from their behaviour.

The Relevance of Philosophy and History to Science

The need to critically evaluate scientific data and claims seems obvious, and it also seems obvious that scientists do this by default. But science students seldom receive formal training in these aspects. The relevance of maths to science is obvious and science students often take compulsory maths courses; but why not an obligatory course on logic or critical thinking? Sophisticated mathematical (or experimental) analyses are wasted if the underlying concepts are flawed. The zoologist Lancelot Hogben wrote, ‘There is a danger of concealing assumptions which have no factual basis behind an impressive façade of flawless algebra.’Footnote ⁵⁹

A grounding in logic and critical thinking would be provided by exposure to philosophy. But the generally negative view of philosophy amongst scientists means that neuroscience students aren’t typically exposed to philosophical aspects (consciousness is an exception). The neuroscientist Semir Zeki claimed philosophers engage in ‘endless and ultimately fruitless discussion’.Footnote ⁶⁰ Some philosophers have agreed: Bertrand Russell said, ‘Philosophy from the earliest times, has made greater claims, and achieved fewer results, than any other branch of learning.’Footnote ⁶¹ But links between science and philosophy have been proposed. Russell said that philosophy required people with ‘scientific training and philosophical interests, unhampered by the traditions of the past,’Footnote ⁶² When Einstein was asked if philosophy should be included in a physics course he said, ‘A knowledge of the historic and philosophical background gives that kind of independence from prejudices of his generation from which most scientists are suffering’, and he claimed this was ‘the mark of distinction between a mere artisan or specialist and a real seeker after truth’.Footnote ⁶³ This wasn’t someone turning philosophical with age: in 1916 Einstein wrote that the best students he had encountered had an interest in philosophy.Footnote ⁶⁴

One potential reason why scientists don’t engage with philosophy is that philosophers use language that is unfamiliar to them. Every field has its own jargon and science is as guilty of this as any other. But some philosophers of neuroscience have introduced new terms for existing concepts. For example, neurobiological criteria for linking molecular and cellular components to behaviour (see Chapter 6) have been repeated using the terms ‘entities’ and ‘activities’ for what I assume are neurons/synapses/molecules and their functional effects, respectively. This may be an attempt to make definitions precise, but it may also be a ‘re-branding’ that tries to suggest novelty when discussing established concepts, something that scientists can also be guilty of. Whatever the reason, the use of multiple terms for seemingly the same aspects can create confusion and hinder discussion. Another issue is that philosophical discussions can appeal to abstractions like zombies that lack consciousness or aliens with silicon rather than carbon-based brains to show that the same functions can be generated despite differences in the underlying components. These philosophical discussions are easy for scientists to ignore because they don’t exist: the philosopher of neuroscience John Bickle put it directly; ‘I’ll worry about brainless yet pained or belief-entertaining aliens and robots as soon as one crosses my path.’Footnote ⁶⁵

Despite these issues, certain philosophical aspects are relevant to any science. This includes what is knowledge, explanation and understanding; how do we ensure our claims are accurate; and how should we use our knowledge. Bennett and Hacker suggest that by analysing the language used to explain neuroscience concepts, philosophy can cut through muddled thinking. Take understanding: we probably all think we know what it is and understanding is frequently claimed in scientific papers. But understanding lacks a formal definition (see Chapter 4). Is discussing the meaning of understanding an example of a pointless philosophical debate, or are scientists claiming something without knowing what it is they are claiming and why? Another example comes from the psychologist Per Holth who claimed we can convert verbs into nouns to generate category mistakes (i.e. assigning a function to something that belongs to another category; for example asking how much the number three weighs).Footnote ⁶⁶ Converting aspects like thinking, feeling, remembering (verbs) into nouns can lead us to think of them as parts of the brain rather than as brain processes. For example, the amygdala has been labelled a ‘fear-centre’. But fear is a mental state that involves multiple brain regions, and while the amygdala promotes freezing behaviour in response to aversive stimuli, it does much more than this, including having a role in positive emotions and socialisation.Footnote ⁶⁷ In addition to being imprecise, language can also be designed to persuade (see the appeal to emotion in Chapter 3). A scientist’s word processor should automatically replace terms like ‘obviously’, ‘unequivocal’ or ‘complete’ with ‘suggested’, ‘putative’, ‘tentative’, ‘hypothesised’. As it won’t, the use of the former terms should invite questions about what is being claimed and why.

This can be all dismissed as philosophical word games, but language can influence thought. An example is to associate functions to neurotransmitters: glutamate is excitatory, GABA inhibitory, serotonin is for mood and dopamine for reward. But these functions don’t simply reflect the neurotransmitter but the receptor the transmitter activates, the cell the receptor is in (e.g. ion concentrations determine whether an ion moves into or out of the cell when a receptor channel opens) and the circuit the cell containing the receptor is in (if a transmitter inhibited neuron x which inhibits inhibitory neuron y, the inhibition of inhibition (disinhibition) equals excitation).

Ascribing functions to neurotransmitters might be a convenient shorthand; glutamate is often excitatory and GABA inhibitory. But claiming serotonin is ‘for’ mood even if its involvement in mood was established is less applicable given that serotonin also influences sensory processing and motor outputs. Likewise, dopamine is associated with reward but isn’t a ‘reward’ transmitter as it also influences light adaptation in the retina. This isn’t simply playing with words, linking functions to neurotransmitters can lead to confusion and potential error. For example, the field of connectomics usefully aims to provide detailed ultrastructural maps of brain regions (see Chapter 8), but some proponents claim to understand function by identifying the neurotransmitter in a presynaptic neuron. Eckstein et al. wrote, ‘The action a neuron has on its downstream targets is determined [my italic] by the neurotransmitters it releases into the extracellular space’.Footnote ⁶⁸ But the action isn’t determined by the neurotransmitter alone, for the reasons outlined above. Another example is the increased levels of the enzyme glutamic acid decarboxylase which synthesises the neurotransmitter GABA, after nervous system injury. The assumption that GABA is inhibitory led to claims that increased inhibition is needed after injury.Footnote ⁶⁹ But again, the effect of GABA depends on the receptor activated and the cell the receptor is in, and after injury GABA can evoke excitation.Footnote ⁷⁰ Even if it did only evoke inhibition, disinhibition means we still cannot assume it functionally has an inhibitory effect. Interventions based on the erroneous view that transmitters determine functions could generate opposite effects to those intended, with obvious negative consequences.

Despite the apparent barrier between science and philosophy, there are moves to bring them together. Some doctoral training programmes have obligatory philosophy courses,Footnote ⁷¹ and scientific meetings are including philosophical discussions. A 2015 meeting on string theory saw physicists and philosophers discussing whether theoretical physics had become divorced from experimental approaches. Ellis and SilkFootnote ⁷² wrote, ‘Faced with difficulties in applying fundamental theories to the observed Universe, some researchers called for a change in how theoretical physics is done. They began to argue – explicitly – that if a theory is sufficiently elegant and explanatory, it need not be tested experimentally, breaking with centuries of philosophical tradition of defining scientific knowledge as empirical.’ This has been termed ‘post-empirical science’. The meeting discussed Karl Popper’s claim that a scientific hypothesis is only valid if it is open to experiment (see Chapter 2). Some theoreticians used the examples of string theory and the multiverse to claim that this requirement was unnecessary, either because these could become testable in the future, or because they are credible hypotheses that are worth pursuing even if they can never be tested experimentally.

These seem worrying trends that take credible hypotheses and pass them off as something more (Ellis and Silk quoted the physicist Sabine Hossenfelder who called post-empirical science an oxymoron). Absence of evidence for alternatives isn’t evidence of their absence and accepting something because it is credible can promote dogma and reduce the motivation to examine the aspect further, making it less likely that alternatives will be found. Theories can be wrong irrespective of how credible they seem, expressed in the quote by nineteenth-century biologist Thomas Huxley, ‘The great tragedy of science – the slaying of a beautiful hypothesis by an ugly fact.’Footnote ⁷³ If we allow post-empiricism in science then why can’t astrologers or homeopathists claim their disciplines are scientific because they seem credible to them or because a scientific basis might eventually be found?

Similar discussions have occurred in the history of science.Footnote ⁷⁴ Brush asked whether history should be written to avoid damaging the image of science by removing examples of irrational behaviour, acrimonious competition, luck, mistakes, or rigid defence of dogma that delayed progress.Footnote ⁷⁵ This would be a ‘Whig’ interpretation of history that would present science as progressing inevitably from ignorance to understanding. But is science undermined by showing its varied inspirations? The neurophysiologist Alan Hodgkin wrote of his Nobel Prize-winning work, ‘I have developed a feeling of guilt about suppressing the part which chance and good fortune played in what now seems to be a rather logical development.’Footnote ⁷⁶

Someone new to a field may consider learning its history is less important than learning its current status and the experimental and analytical techniques they will need to use. But history shows how a scientific field has developed. While a simpler story might be easier to follow, the history of a field can highlight potential pitfalls, reveal forgotten insights that may help advance understanding, or highlight that an anomaly holding a field back reflects an untested assumption that has prevented better explanations from being considered. A realistic history of science also better prepares students for a scientific career in which they are unlikely to enjoy the continuous progression and co-operation that a Whig history of science might suggest was the norm. The physiologist Emil du Bois-Reymond claimed that ‘the best way of communicating science may be narrating its history,’Footnote ⁷⁷ and Peter Medawar said a young scientist ‘should certainly try to identify the origin and growth of current opinions … everything that a scientist does is a function of what others have done before: the past is embodied in every new conception’.Footnote ⁷⁸

We should be careful not to facilitate anti-science views by promoting post-empirical science or fictional histories (a sceptic could argue if scientists make up the history of their field, what else are they making up?). An honest account of science would highlight that error, chance and uncertainty are inevitable. Claude Bernard wrote, ‘Even mistaken hypotheses and theories are of use in leading to discoveries. The alchemists founded chemistry by pursuing chimerical problems and theories which were false … It seems, indeed, a necessary weakness of our minds to be able to reach truth only across a multitude of error and obstacles.’Footnote ⁷⁹ We can also admit the role of luck – for example, in the discovery of penicillin by Fleming, the role of insulin in diabetes by Minkowski and X-rays by Roentgen – but highlight that even chance discoveries needed scientific insight to recognise their importance. The Hungarian biochemist Albert Szent-Gyorgyi said, ‘Discovery consists of seeing what everybody has seen and thinking what nobody has thought’; and Louis Pasteur said, ‘In the fields of observation, chance favors only the prepared mind.’Footnote ⁸⁰ We can also admit influences other than logic and rationality. Both Michael Faraday and Albert Einstein claimed to have been poor students, but both could imagine aspects that others couldn’t. Einstein said, ‘Imagination is more important than knowledge. For knowledge is limited, whereas imagination embraces the entire world, stimulating progress … It is, strictly speaking, a real factor in scientific research.’Footnote ⁸¹

Outlining our limited understanding of brain function, why it goes wrong and how to correct it, or the hyperbole of neuroscience claims to fully conscious brain emulations, memory downloads and telepathy, can lead to accusations of pessimism or nihilism that damages the field. But considering what we can and cannot say with confidence doesn’t seem pessimistic or nihilistic, but realistic. Pessimism to me arises when erroneous claims and hyperbole are uncritically accepted as explanations or understanding, if debate is discouraged or blocked and if issues are hidden to promote an individual or a field, because no problem will be solved by ignoring it. Erroneous claims ultimately delay progress because they first have to be identified and negated before the correct explanation can be approached. We should be able to treat neuroscience claims critically and dismiss hyperbole while still recognising the field’s significant achievements.

We don’t need to claim the brain is more complex than any other object in the universe (how would we know?), it simply is complex. We know a lot, but not as much as we would like to or need to (we are data rich but theory light). The psychologist Gary Marcus claimed ‘neuroscience has yet to find its Newton, let alone its Einstein,’Footnote ⁸² which would leave it about 400 years behind physics. Although we have asked questions about the brain for millennia, its scientific analysis is relatively recent. Scientific advance, even though often linked to a key experiment or insight, typically reflects a prolonged period of prior analyses: the structure of DNA, the biochemical pathways for ATP synthesis and electrical signalling in the nervous system, all aspects included in school textbooks, took decades to establish. Slow, incremental steps are the norm in science even when the final step is associated with a ‘eureka’ moment. We cannot know if the work we do as scientists will ultimately make a significant contribution. Even if everything we did was always accurate and reliable, insight occurs when the information we obtain is applied to a relevant concept. The classic example is silicon: the significance of the incremental understanding of its basic chemical and physical properties doesn’t need to be stated.Footnote ⁸³

Science is claimed to differ from political or religious ideology because of its ‘intellectual flexibility, a graceful (albeit sometimes grudging) acceptance of the need to change our minds as our understanding of the world evolves’.Footnote ⁸⁴ This is the ideal, but change may not come easily. Even with the best intent of accuracy, experimenting on biological systems requires addressing multiple variables that we are either unaware of or we can’t control, and seemingly minor differences in an initial state can lead to marked differences in experimental results and interpretations. Various cognitive biases and logical fallacies can also affect our conclusions (see Chapter 3), as can the pressures of scientific career structures that require publishing in terms of both volume and impact (Chapter 10). I will give examples in subsequent chapters, but an example to end with is the resistance by the nineteenth-century medical establishment to the demonstration by the Hungarian physician Ignaz Semmelweis that often-fatal puerperal fever after childbirth was dramatically reduced if obstetricians simply washed their hands (or if women gave birth in the street, which many did to avoid clinics where the incidence of puerperal fever was high). Semmelweis was ignored and then vilified by the medical establishment, leading to a nervous breakdown that saw him committed to an asylum where he died soon after. A recent, albeit less dramatic, neuroscience example is adult neurogenesis (see Chapter 2). In both cases change occurred, evidencing intellectual flexibility, but resistance to change made the process slow and inefficient. In the next chapter I will discuss why resistance to change can be expected and, to some extent, welcomed in science.