The Neuroscience of Intelligence

by Richard J Haier

  None of the findings reported so far are advanced enough to consider actual genetic engineering to produce highly intelligent children. There is a recent noteworthy development in genetic engineering technology, however, with implications for enhancement possibilities. A new method for editing the human genome is called CRISPR/Cas9 (Clustered Regularly Interspaced Short Palindromic Repeats/Cas genes). I don’t understand the name either, but this method uses bacteria to edit the genome of living cells by making changes to targeted genes (Sander & Joung, 2014). It is noteworthy because many researchers can apply this method routinely so that editing the entire human genome is possible as a mainstream activity. Once genes for intelligence and how they function are identified, this kind of technology could provide the means for enhancement on a large scale. Perhaps that is why the name of the method was chosen to be incomprehensible to most of us. Keep this one on your radar too.

  Most of this chapter is about what does not work to enhance intelligence. It is fair to say that education and cognitive approaches have made little demonstrable progress after many years of concerted efforts and neuroscience approaches are relatively nascent. We should not be discouraged, just as we are not discouraged that the hunt for intelligence genes has progressed slowly. The brain is complex and its secrets are not easily revealed. All science is technology-driven and intelligence research is no exception. As discussed in this chapter and previous ones, there are exciting possibilities for enhancing intelligence based on new brain technologies and new information about brain structure, function, and development. From my perspective of nearly 45 years in the field, the pace of discovery is quickening. There is no clear roadmap for the future, but the next and final chapter will present some neuroscience perspectives on emerging approaches for learning even more about intelligence and the brain.

  Chapter 5 Summary

  Despite many claims, there is yet no way to increase any intelligence factor that survives independent replication and creates a compelling weight of evidence.

  Studies that have made claims of enhancement have serious flaws including “teaching to the test,” generalizing from small samples, and treating small score changes on single tests as indications of large changes in underlying intelligence factors.

  Psychoactive drugs and various non-drug methods of stimulating the brain may have potential for cognitive enhancement of attention, learning, and memory, but there is no weight of evidence yet that these methods enhance intelligence.

  Ultimately, enhancement may depend on not only finding specific genes related to intelligence but also on the harder problem of understanding how those genes function on a molecular level, including epigenetic influences.

  Review Questions

  Why is the “weight of evidence” concept especially important for claims about enhancing intelligence?

  What are three examples of research findings which claimed sizeable increases in IQ that proved incorrect?

  Explain the concepts of “transfer” and “independent replication.”

  What are five methods of brain stimulation that may influence cognition?

  What are six ethical issues concerning the use of drugs for cognitive enhancement?

  Further Reading

  “Cognitive enhancement” (Farah et al., 2014). This is a comprehensive discussion of enhancement issues.

  “Increased intelligence is a myth (so far)” (Haier, 2014). Explains why intelligence test score increases do not mean intelligence has increased.

  Chapter Six

  As Neuroscience Advances, What’s Next for Intelligence Research?

  We choose to go to the moon and do the other things … not because they are easy, but because they are hard.

  (President John F. Kennedy, speech at Rice University, September 12, 1962)

  The remarkable thing is that although basic research does not begin with a particular practical goal, when you look at the results over the years, it ends up being one of the most practical things government does.

  (President Ronald Reagan, radio address, April 1, 1988)

  Without a doubt, this is the most important, most wondrous map ever produced by human kind.

  (President Bill Clinton, remarks on completion of the first survey of the entire Human Genome Project, June 26, 2000)

  As humans we can identify galaxies light years away, we can study particles smaller than an atom, but we still haven’t unlocked the mystery of the three pounds of matter between our ears.

  (President Barack Obama, statement introducing the Federal Human Brain Initiative, April 2, 2013)

  Learning Objectives

  What is chronometrics and why is it an advance over psychometrics?

  How does the study of memory and super-memory inform research on intelligence?

  How does research on animals provide insights about neurons and intelligence?

  How does neuroscientific understanding of brain circuits advance building intelligent machines?

  Given problems of definition, how can there be a neuroscience of consciousness and creativity?

  Why might social–economic status (SES) and intelligence be confounded on the neural level?

  Introduction

  Paradoxically, in any area of scientific inquiry, the more we learn, the more we do not understand. An answer to one question often leads to a new question never before formulated. Advances depend on our intellect and imagination to make sense of new empirical observations obtained from creative methods and technologies that are constantly improving to provide new kinds of data. Think about the experimental validation of the Standard Model in particle physics as a result of observations made with multibillion-dollar accelerators. These huge, worldwide efforts also revealed new mysteries like dark energy that cannot be solved by existing methods so new ones must be invented. Each generation of researchers builds on the recent past and extends into the near future. The early researchers who studied rudimentary language and perceptual deficits in patients with brain damage could not imagine the neuroscience tools now available to address questions about intelligence and the brain. All the advances in genetics and imaging methods and the potential for understanding and perhaps enhancing intelligence discussed so far in previous chapters are just the beginning. There is more to come, but the pace depends on whether there is a commitment for generous funding that is dispensed wisely. A focus on research that will likely yield practical results is not necessarily wise, as history shows many examples of unforeseen major benefits from seemingly arcane basic research. It is nearly impossible to imagine, but what if a country ignored space exploration and announced its major scientific goal was to achieve the capability to increase every citizen’s g-factor by a standard deviation? By the end of this chapter, you might not think this is so impossible.

  In every area of science, each stage of progress becomes more expensive and complex to conduct logistically and it becomes more complex to interpret results. For neuroimaging, CAT scans are more complex than X-rays; structural MRI is more complex than CAT. PET is more complex than EEG; functional MRI is more complex than structural MRI; MRI spectroscopy and diffusion tensor imaging (DTI) are more complex than structural or functional MRI; MEG is more complex than MRI. Each new technology provides better spatial and temporal resolutions and amasses bigger and bigger data files that require more advanced computer power for processing and analyses. MRI shows brain tissue in millimeters, but even this is far too big for showing individual neurons or synapses. MEG shows brain activity changes every millisecond, but this is far too slow to show nanosecond neurochemical events in the synapse. Neuroscience techniques are available to study the brain at the level of single neurons and synapses, so it is not beyond imagination that these techniques can be applied to questions about intelligence. Advances in the intelligence field likely will come from the integration of findings from basic research on clinical brain disorders, aging, and normal cognitive processes like learning, memory, and attention from both animal and human s
tudies that expose events smaller and smaller, faster and faster, and deeper and deeper in the brain.

  This chapter will highlight six developing lines of inquiry relevant to intelligence. Before discussing them, here is a brief recap of three main points developed in the previous chapters. (1) Based on the weight of evidence, intelligence is something that can be defined, measured and studied scientifically, especially the g-factor, which correlates to many real-world outcomes, brain structure and function, and has a strong genetic basis. (2) Neuroimaging research is beginning to identify specific brain characteristics related to intelligence differences among people, and genetic research is beginning to identify specific genes related to intelligence. These advances, driven by technology, are moving intelligence research in a more neuroscience direction. (3) How brain characteristics related to intelligence develop from genetic, biological and environmental factors, and their interactions, is not yet understood. But once we have a better understanding of how these factors work in the brain, we should be able to manipulate them to increase intelligence either to close any gaps among groups or raise everyone, perhaps dramatically. Building on these three points and moving forward, here are six exciting areas to watch for progress, each in its own section.

  6.1 From Psychometric Testing to Chronometric Testing

  On one side of the equation that links genetic and neuroimaging data to intelligence, we have the most up-to-date multimillion-dollar equipment and teams of specialists to collect and analyze complex data sets. On the other side of the equation, we have a psychometric test score, often from a single test that costs a few dollars. This is quite a mismatch, or more accurately a chasm. Decades ago, the earliest imaging studies of intelligence and the earliest quantitative and molecular genetic studies of intelligence used the same intelligence tests still used today. To advance the field, the study of intelligence can no longer be limited to psychometric test scores. As noted in Chapter 1, a sophisticated measurement of intelligence is badly needed to match the sophisticated genetic and neuroimaging assessments widely available. At minimum, a latent variable approach that extracts a factor from a battery of tests is required. The optimal assessment of intelligence will require a ratio scale, as noted in Chapter 1.

  Let’s review why this is so using a new example. Suppose you have an intervention that is designed to increase happiness (pick whatever intervention you like). You measure happiness by asking participants to rate their happiness on a scale from 1 to 10, where a 10 represents the most happiness. You find an average happiness score of 4 for a group before the intervention. After the intervention, the group average has increased to 8. If you are naïve about measuring constructs like happiness, you might conclude that your intervention resulted in people becoming twice as happy based on the change from 4 to 8. You would be wrong to conclude this. Your happiness scale is an interval scale where points are not equivalent and each person has a subjective idea of what 4 or 8 means. Eight on an interval scale is not literally 2 × 4. Eight pounds, however, is literally 2 × 4 pounds because pounds are a ratio scale that is bounded by an actual zero point of no weight. A pound of bricks weighs the same as a pound of feathers. A pound is a pound irrespective of what is being measured.

  Intelligence test scores, like all measures of happiness, are all on interval scales. Your score has meaning only relative to other people, typically expressed as a percentile. If you are at the 95th percentile, how much more intelligent are you than someone at the 90th percentile? You are not 5% more intelligent. We do not have a measure of intelligence as a quantity. In Chapter 4, we discussed whether intelligence could be defined by quantifying brain characteristics like the amount of gray matter, thickness of the cortex, connectivity of networks, or the integrity of white matter. These are all potential ratio scales, but imaging is not a practical basis for wide use as an intelligence test in most settings. Another way to create a ratio measurement of intelligence depends on the measurement of time (8 seconds is literally twice 4 seconds). The concept is to create a standard battery of mental tests where the time it takes to arrive at an answer is the basis for the measurement rather than the number of correct answers. Intelligence could then be defined as speed of information processing during a standard set of test items. A person who had an average time of 4 seconds on a battery of information-processing test items would be literally twice as fast as a person with an average of 8 seconds. The validity of information-processing speed as an alternative definition of intelligence would require research establishing what this metric might predict in terms of academic or other achievement. In fact, a considerable body of research like this already exists.

  In his last book before he died, Arthur Jensen summarized this research and considered the technical obstacles to overcome for developing a new kind of intelligence test based on information-processing time (Jensen, 2006). He called this approach “chronometrics.” A chronometric testing apparatus is currently under development and is described in Textbox 6.1. Note the test items are much different than those of psychometric tests. For example, one test has a semicircle of eight buttons that can light up. For each trial three buttons light up at the same time. As fast as you can, you press the lighted button farthest away from the other two lit buttons. After a series of trials on this test, few people make many mistakes and test results are measured in units of time, so chronometric scores are on ratio scales. If research supports the validity and reliability of the chronometric approach by establishing correlations with intelligence test scores, its use in future genetic and neuroimaging studies could narrow the sophistication-of-measurement gap. Jensen expressed the optimistic view that chronometric approaches could elevate intelligence research to a natural science. Combined with other neuroscience approaches, the pace of discovery would surely increase with this kind of measurement.

  It may even be possible to define intelligence by brain characteristics such as speed of information processing or the amount of gray matter tissue in certain areas. The advantage of such definitions would be that they are quantitative on a ratio scale. Imagine that your information-processing speed on a standard test battery is twice as fast as someone else’s. Whether this might predict something about your future academic success or other variables better than an IQ score is an open question for empirical study.

  Textbox 6.1: Chronometric assessment of intelligence

  As proposed by Jensen (2006), mental chronometry is based on two fundamental concepts. The first is that the time it takes to make a decision is a measure of brain processing speed. This is often referred to as reaction time (RT) or response time. RT studies have a long history in psychology, going back more than 100 years. Many cognitive tasks have been used in RT studies. Often they are called elementary cognitive tasks (ECTs). One of the most replicated findings is that RT increases with task complexity. Another core finding is that people with faster RTs generally have higher IQ scores. Therefore, the measurement of RT could be used to measure intelligence and RT is an especially attractive metric because time is a ratio scale. RT for most ECTs is measured in milliseconds or seconds, depending on the task. The second fundamental concept is standardization. Different researchers often do RT studies using different devices. This lack of uniformity introduces method variance that confounds the RT assessment of individuals and makes it difficult to compare studies or to combine data from different studies into a large data set. Jensen proposed building a standard device to test RT to a standard set of diverse ECTs. Jensen believed that the combination of RT measures with a standardized method for testing ECTs would advance the study of intelligence beyond the limitations of psychometric tests like the WAIS. The Institute of Mental Chronometry (IMC) was founded and funded by Jensen to develop and disseminate such a device. At this time a prototype is under development. The device has a combined display screen and a button response panel with eight buttons arranged in a semi-circle and a home button below the semi-circle. A keyboard and mouse can be attached so the investigator
can set parameters for any experiment.

  For example, one ECT involves the eight buttons. To start a trial, the person being tested presses a finger on the home button, holding it down. Three of the eight buttons then light up simultaneously. One of them will be further away from the other two. As quickly as possible, the person releases the home button and presses the lighted button furthest away from the other two (see Animation 6.1 on this book’s website for a demonstration, www.cambridge.org/us/academic/subjects/psychology/cognition/neuroscience-intelligence). This is called an “Odd Man Out task”. After a series of such trials, a person’s average RT is computed. A different ECT requires the person to memorize a string of numbers (or letters or shapes) after seeing them for a brief period on the display screen. Then a target number (or letter or shape) appears and if the target was in the string memorized, the designated yes button is pressed. If the target is not in the string, the no button is pressed. As the trials continue, the string gets longer so more memory is scanned in order to decide yes or no. This increases RT and higher-IQ people generally scan memory faster than lower-IQ people. Another ECT shows two words on the screen simultaneously. If they are synonyms, one button is pressed; if not, another button is pressed. There are many variations of these tasks and many other ECTs. Research will establish which ECTs will generate RTs that, in combination, make a good battery to assess intelligence. There are many technical issues to resolve. There is a long road ahead for this research before mental chronometry might replace psychometric tests of intelligence. At the end of his book, Jensen concluded, “… chronometry provides the behavioral and brain sciences with a universal absolute [ratio] scale for obtaining highly sensitive and frequently repeatable measurements of an individual’s performance on specially devised cognitive tasks. Its time has come. Let us get to work!” (p. 246). This method of assessing intelligence could establish actual changes due to any kind of proposed enhancement in a before and after research design. The sophistication of this method for measuring intelligence would diminish the gap with sophisticated genetic and neuroimaging methods.