bioliteracy / biofundamentals

Making education matter in higher education

It may seem self-evident that providing an effective education, the type of educational experiences that lead to a useful bachelors degree and serve as the foundation for life-long learning and growth, should be a prime aspirational driver of Colleges and Universities (1). We might even expect that various academic departments would compete with one another to excel in the quality and effectiveness of their educational outcomes; they certainly compete to enhance their research reputations, a competition that is, at least in part, responsible for the retention of faculty, even those who stray from an ethical path. Institutions compete to lure research stars away from one another, often offering substantial pay raises and research support (“Recruiting or academic poaching?”). Yet, my own experience is that a department’s performance in undergraduate educational outcomes never figures when departments compete for institutional resources, such as supporting students, hiring new faculty, or obtaining necessary technical resources (2).

I know of no example (and would be glad to hear of any) of a University hiring a professor based primarily on their effectiveness as an instructor (3).

In my last post, I suggested that increasing the emphasis on measures of departments’ educational effectiveness could help rebalance the importance of educational and research reputations, and perhaps incentivize institutions to be more consistent in enforcing ethical rules involving research malpractice and the abuse of students, both sexual and professional. Imagine if administrators (Deans and Provosts and such) were to withhold resources from departments that are performing below acceptable and competitive norms in terms of undergraduate educational outcomes?

Outsourced teaching: motives, means and impacts

Sadly, as it is, and particularly in many science departments, undergraduate educational outcomes have little if any impact on the perceived status of a department, as articulated by campus administrators. The result is that faculty are not incentivized to, and so rarely seriously consider the effectiveness of their department’s course requirements, a discussion that would of necessity include evaluating whether a course’s learning goals are coherent and realistic, whether the course is delivered effectively, whether it engages students (or is deemed irrelevant), and whether students’ achieve the desired learning outcomes, in terms of knowledge and skills achieved, including the ability to apply that knowledge effectively to new situations. Departments, particularly research focussed (dependent) departments, often have faculty with low teaching loads, a situation that incentivizes the “outsourcing” of key aspects of their educational responsibilities. Such outsourcing comes in two distinct forms, the first is requiring majors to take courses offered by other departments, even if such courses are not well designed, delivered, or (in the worst cases) relevant to the major. A classic example is to require molecular biology students to take macroscopic physics or conventional calculus courses, without regard to whether the materials presented in these courses is ever used within the major or the discipline. Expecting a student majoring in the life sciences to embrace a course that (often rightly) seems irrelevant to their discipline can alienate a student, and poses an unnecessary obstacle to student success, rather than providing students with needed knowledge and skills. Generally, the incentives necessary to generate a relevant course, for example, a molecular level physics course that would engage molecular biology students, are simply not there. A version of this situation is to require courses that are poorly designed or delivered (general chemistry is often used as the poster child for such a course). These are courses that have high failure rates, sometimes justified in terms of “necessary rigor” when in fact better course design could (and has) resulted in lower failure rates and improved learning outcomes. In addition, there are perverse incentives associated with requiring “weed out” courses offered by other departments, as they reduce the number of courses a department’s faculty needs to teach, and can lead to fewer students proceeding into upper division courses.

The second type of outsourcing involves excusing tenure track faculty from teaching introductory courses, and having them replaced by lower paid instructors or lecturers. Independently of whether instructors, lecturers, or tenure track professors make for better teaching, replacing faculty with instructors sends an implicit message to students. At the same time, the freedom of instructors/lecturers to adopt an effective (socratic) approach to teaching is often severely constrained; common exams can force classes to move in lock step, independently of whether that pace is optimal for student engagement and learning. Generally, instructors/lecturers do not have the freedom to adjust what they teach, to modify the emphasis and time they spend on specific topics in response to their students’ needs. How an instructor instructs their students suffers when teachers do not have the freedom to customize their interactions with students in response to where they are intellectually. This is particularly detrimental in the case of underrepresented or underprepared students. Generally, a flexible and adaptive approach to instruction (including ancillary classes on how to cope with college: see An alternative to remedial college classes gets results) can address many issues, and bring the majority of students to a level of competence, whereas tracking students into remedial classes can succeed in driving them out of a major or college (see Colleges Reinvent Classes to Keep More Students in Science and Redesigning a Large-Enrollment Introductory Biology Course and Does Remediation Work for All Students? )

How to address this imbalance, how can we reset the pecking order so that effective educational efforts actually matter to a department?

My (modest) suggestion is to base departmental rewards on objective measures of educational effectiveness. And by rewards I mean both at the level of individuals (salary and status) as well as support for graduate students, faculty positions, start up funds, etc. What if, for example, faculty in departments that excel at educating their students received a teaching bonus, or if the number of graduate students within a department supported by the institution was determined not by the number of classes these graduate students taught (courses that might not be particularly effective or engaging) but rather by a departments’ undergraduate educational effectiveness, as measured by retention, time to degree, and learning outcomes (see below)? The result could well be a drive within a department to improve course and curricular effectiveness to maximize education-linked rewards. Given that laboratory courses, the courses most often taught by science graduate students, are multi-hour schedule disrupting events, of limited demonstrable educational effectiveness, that complicate student course scheduling, removing requirements for lab courses deemed unnecessary (or generating more effective versions), would be actively rewarded (of course, sanctions for continuing to offer ineffective courses would also be useful, but politically more problematic.)

A similar situation applies when a biology department requires its majors to take 5 credit hour physics or chemistry courses. Currently it is “easy” for a department to require its students to take such courses without critically evaluating whether they are “worth it”, educationally. Imagine how a department’s choices of required courses would change if the impact of high failure rates (which I would argue is a proxy for poorly designed and delivered courses) directly impacted the rewards reaped by a department. There would be an incentive to look critically at such courses, to determine whether they are necessary and if so, well designed and delivered. Departments would serve their own interests if they invested in the development of courses that better served their disciplinary goals, courses likely to engage their students’ interests.

So how do we measure a department’s educational efficacy?

There are three obvious metrics: i) retention of students as majors (or in the case of “service courses” for non-majors, whether students master what it is the course claims to teach); ii) time to degree (and by that I mean the percentage of students who graduate in 4 years, rather than the 6 year time point reported in response to federal regulations (six year graduation rate | background on graduation rates); and iii) objective measures of student learning outcomes attained and skills achieved. The first two are easy, Universities already know these numbers. Moreover they are directly influenced by degree requirements – requiring students to take boring and/or apparently irrelevant courses serves to drive a subset of students out of a major. By making courses relevant and engaging, more students can be retained in a degree program. At the same time, thoughtful course design can help students pass through even the most rigorous (difficult) of such courses. The third, learning outcomes, is significantly more challenging to measure, since universal metrics are (largely) missing or superficial. A few disciplines, such as chemistry, support standardized assessments, although one could argue with what such assessments measure. Nevertheless, meaningful outcomes measures are necessary, in much the same way that Law and Medical boards and the Fundamentals of Engineering exam serve to help insure (although they do not guarantee) the competence of practitioners. One could imagine using parts of standardized exams, such as discipline specific GRE exams, to generate outcomes metrics, although more informative assessment instruments would clearly be preferable. The initiative in this area could be taken by professional societies, college consortia (such as the AAU), and research foundations, as a critical driver for education reform, increased effectiveness, and improved cost-benefit outcomes, something that could help address the growing income inequality in our country and make success in higher education an important factor contributing to an institution’s reputation.

A footnote or two…

1. My comments are primarily focused on research universities, since that is where my experience lies; these are, of course, the majority of the largest universities (in a student population sense).

2. Although my experience is limited, having spent my professorial career at a single institution, conversations with others leads me to conclude that it is not unique.

3. The one obvious exception would be the hiring of coaches of sports teams, since their success in teaching (coaching) is more directly discernible and impactful on institutional finances and reputation).

minor edits – 16 March 2020

Balancing research prestige, human decency, and educational outcomes.

Or why do academic institutions shield predators? Many working scientists, particularly those early in their careers or those oblivious to practical realities, maintain an idealistic view of the scientific enterprise. They see science as driven by curious, passionate, and skeptical scholars, working to build an increasingly accurate and all encompassing understanding of the material world and the various phenomena associated with it, ranging from the origins of the universe and the Earth to the development of the brain and the emergence of consciousness and self-consciousness (1). At the same time, the discipline of science can be difficult to maintain (see PLoS post: The pernicious effects of disrespecting the constraints of science). Scientific research relies on understanding what people have already discovered and established to be true; all too often, exploring the literature associated with a topic can reveal that one’s brilliant and totally novel “first of its kind” or “first to show” observation or idea is only a confirmation or a modest extension of someone else’s previous discovery. That is the nature of the scientific enterprise, and a major reason why significant new discoveries are rare and why graduate students’ Ph.D. theses can take years to complete.

Acting to oppose a rigorous scholarly approach are the real life pressures faced by working scientists: a competitive landscape in which only novel observations get rewarded by research grants and various forms of fame or notoriety in one’s field, including a tenure-track or tenured academic position. Such pressures encourage one to distort the significance or novelty of one’s accomplishments; such exaggerations are tacitly encouraged by the editors of high profile journals (e.g. Nature, Science) who seek to publish “high impact” claims, such as the claim for “Arsenic-life” (see link). As a recent and prosaic example, consider a paper that claims in its title that “Dietary Restriction and AMPK Increase Lifespan via Mitochondrial Network and Peroxisome Remodeling” (link), without mentioning (in the title) the rather significant fact that the effect was observed in the nematode C. elegans, whose lifespan is typically between 300 to 500 hours and which displays a trait not found in humans (and other vertebrates), namely the ability to assume a highly specialized “dauer” state that can survive hostile environmental conditions for months. Is the work wrong or insignificant? Certainly not, but it is presented to the unwary (through the Harvard Gazette under the title, “In pursuit of healthy aging: Harvard study shows how intermittent fasting and manipulating mitochondrial networks may increase lifespan,” with the clear implication that people, including Harvard alumni, might want to consider the adequacy of their retirement investments

Such pleas for attention are generally quickly placed in context and their significance evaluated, at least within the scientific community – although many go on to stimulate the economic health of the nutritional supplement industry. Lower level claims often go unchallenged, just part of the incessant buzz associated with pleas for attention in our excessively distracted society (see link). Given the reward structure of the modern scientific enterprise, the proliferation of such claims is not surprising. Even “staid” academics seek attention well beyond the immediate significance of their (tax-payer funded) observations. Unfortunately, the explosively expanding size of the scientific enterprise makes policing such transgressions (generally through peer review or replication) difficult or impossible, at least in the short term.

The hype and exaggeration associated with some scientific claims for attention are not the most distressing aspect of the quest for “reputation.” Rather, there are growing number of revelations of academic institutions protecting those guilty of abusing their dependent colleagues. These reflect how scientific research teams are organized. Most scientific studies involve groups of people working with one another, generating data, testing ideas, and eventually publishing their observations and conclusions, and speculating on their broader implications.

Research groups can vary greatly in size. In some areas, they involve isolated individuals, whether thinkers (theorists) or naturalists, in the mode of Darwin and Wallace. In other cases, these are larger and include senior researchers, post-doctoral fellows, graduate students, technicians, undergraduates, and even high school students. Such research groups can range from the small (2 to 3 people) to the significantly larger (~20-50 people); the largest of such groups are associated mega-projects, such as the human genome project and the Large Hadron Collider-based search for the Higgs boson (see: Physics paper sets record with more than 5,000 authors). A look at this site [link] describing the human genome project reflects two aspects of such mega-science: 1) while many thousands of people were involved [see Initial sequencing and analysis of the human genome], generally only the “big names” are singled out for valorization (e.g., receiving a Nobel Prize). That said, there would be little or no progress without general scientific community that evaluates and extends ideas and observations. In this context, “lead investigators” are charged primarily with securing the funds needed to mobilize such groups, convincing funders that the work is significant; it is members of the group that work out the technical details and enable the project to succeed.

As with many such social groups, there are systems in play that serve to establish the status of the individuals involved – something necessary (apparently) in a system in which individuals compete for jobs, positions, and resources. Generally, one’s status is established through recommendations from others in the field, often the senior member(s) of one’s research group or the (generally small) group of senior scientists who work in the same or a closely related area. The importance of professional status is particularly critical in academia, where the number of senior (e.g. tenured or tenure-track professorships) is limited. The result is a system that is increasingly susceptible to the formation of clubs, membership in which is often determined by who knows who, rather than who has done what (see Steve McKnight’s “The curse of committees and clubs”). Over time, scientific social status translates into who is considered productive, important, trustworthy, or (using an oft-misused term) brilliant. Achieving status can mean putting up with abusive and unwanted behaviors (particularly sexual). Examples of this behavior have recently been emerging with increasing frequency (which has been extensively described elsewhere: see Confronting Sexual Harassment in Science; More universities must confront sexual harassment; What’s to be done about the numerous reports of faculty misconduct dating back years and even decades?; Academia needs to confront sexism; and The Trouble With Girls’: The Enduring Sexism in Science).

So why is abusive behavior tolerated? One might argue that this reflects humans’ current and historical obsession with “stars,” pharaohs, kings, and dictators as isolated geniuses who make things work. Perhaps the most visible example of such abused scientists (although there are in fact many others : see History’s Most Overlooked Scientists) is Rosalind Franklin, whose data was essential to solving the structure of double stranded DNA, yet whose contributions were consistently and systematically minimized, a clear example of sexual marginalization. In this light, many is the technician who got an experiment to “work,” leading to their research supervisor’s being awarded the prizes associated with the breakthrough (2).

Amplifying the star effect is the role of research status at the institutional level; an institution’s academic ranking is often based upon the presence of faculty “stars.” Perhaps surprisingly to those outside of academia, an institution’s research status, as reflected in the number of stars on staff, often trumps its educational effectiveness, particularly with undergraduates, that is the people who pay the bulk of the institution’s running costs. In this light, it is not surprising that research stars who display various abusive behavior (often to women) are shielded by institutions from public censure.

So what is to be done? My own modest proposal (to be described in more detail in a later post) is to increase the emphasis on institution’s (and departments within institutions) effectiveness at undergraduate educational success. This would provide a counter-balancing force that could (might?) place research status in a more realistic context.

a footnote or two:

on the assumption that there is nothing but a material world.
Although I am no star, I would acknowledge Joe Dent, who worked out the whole-mount immunocytochemical methods that we have used extensively in our work over the years).
Thanks to Becky for editorial comments as well as a dramatic reading!

Humanized mice & porcinized people

Updates: 12 January 2022

7 December 2020: US FDA declares genetically modified pork ‘safe to eat‘

A practical benefit, from a scientific and medical perspective, of the evolutionary unity of life (link) are the molecular and cellular similarities between different types of organisms. Even though humans and bacteria diverged more than 2 billion years ago (give or take), the molecular level conservation of key systems makes it possible for human insulin to be synthesized in and secreted by bacteria and pig-derived heart valves to be used to replace defective human heart valves (see link). Similarly, while mice, pigs, and people are clearly different from one another in important ways they have, essentially, all of the same body parts. Such underlying similarities raise interesting experimental and therapeutic possibilities.

A (now) classic way to study the phenotypic effects of human-specific versions of genes is to introduce these changes into a model organism, such as a mouse (for a review of human brain-specific human genes – see link). A example of such a study involves the gene that encodes the protein foxp2, a protein involved in the regulation of gene expression (a transcription factor). The human foxp2 protein differs from the foxp2 protein in other primates at two positions; foxP2 evolution these two amio acid changes alter the activity of the human protein, that is the ensemble of genes that it regulates. That foxp2 has an important role in humans was revealed through studies of individuals in a family that displayed a severe language disorder linked to a mutation that disrupts the function of the foxp2 protein. Individuals carrying this mutant foxp2 allele display speech apraxia, a “severe impairment in the selection and sequencing of fine oral and facial movements, the ability to break up words into their constituent phonemes, and the production and comprehension of word inflections and syntax” (cited in Bae et al, 2015). Male mice that carry this foxp2 mutation display changes in the “song” that they sing to female mice (1), while mice carrying a humanized form of foxp2 display changes in “dopamine levels, dendrite morphology, gene expression and synaptic plasticity” in a subset of CNS neurons (2). While there are many differences between mice and humans, such studies suggest that changes in foxp2 played a role in human evolution, and human speech in particular.

Another way to study the role of human genes using mouse as a model system is to generate what are known as chimeras, named after the creature in Greek mythology composed of parts of multiple organisms. A couple of years ago, Goldman and colleagues (3) reported that human glial progenitor cells could, when introduced into immune-compromised mice (to circumvent tissue rejection), displaced the mouse’s own glia, replacing them with human glia cells. iPSC transplant Glial cells are the major non-neuronal component of the central nervous system. Once thought of as passive “support” cells, it is now clear that the two major types of glia, known as astrocytes and oligodendrocytes, play a number of important roles in neural functioning [back track post]. In their early studies, they found that the neurological defects associated with the shaker mutation, a mutation that disrupts the normal behavior of oligodendrocytes, could be rescued by the implantation of normal human glial progenitor cells (hGPCs)(4). Such studies confirmed what was already known, that the shaker mutation disrupts the normal function of myelin, the insulating structure around axons that dramatically speeds the rate at which neuronal signals (action potentials) move down the axons and activate the links between neurons (synapses). In the central nervous system, myelin is produced by oligodendrocytes as they ensheath neuronal axons. Human oligodendrocytes derived from hGPCs displaced the mouse’s mutation carrying oligodendrocytes and rescued the shaker mouse’s mutation-associated neurological defect.

golgi staining- diagram Subsequently, Goldman and associates used a variant of this approach to introduce hGPCs (derived from human embryonic stem cells) carrying either a normal or mutant version of the Huntingtin protein, a protein associated with the severe neural disease Huntington’s chorea (OMIM: 143100)(5). Their studies strongly support a model that locates defects associated with human Huntington’s disease to defects in glia. This same research group has generated hGPCs from patient-derived, induced pluripotent stem cells (patient-derived HiPSCs). In this case, the patients had been diagnosed with childhood-onset schizophrenia (SCZ) [link](6). Skin biopsies were taken from both normal and children diagnosed with SCZ; fibroblasts were isolated, and reprogrammed to form human iPSCs. These iPSCs were treated so that they formed hGPCs that were then injected into mice to generate chimeric (human glial/mouse neuronal) animals. The authors reported systematic differences in the effects of control and SCZ-derived hGPCs; “SCZ glial mice showed reduced prepulse inhibition and abnormal behavior, including excessive anxiety, antisocial traits, and disturbed sleep”, a result that suggests that defects in glial behavior underlie some aspects of the human SCZ phenotype.

The use of human glia chimeric mice provides a powerful research tool for examining the molecular and cellular bases for a subset of human neurological disorders. Does it raise a question of making mice more human? Not for me, but perhaps I do not appreciate the more subtle philosophical and ethical issues involved. The mice are still clearly mice, most of their nervous systems are composed of mouse cells, and the overall morphology, size, composition, and organization of their central nervous systems are mouse-derived and mouse-like. The situation becomes rather more complex and potentially therapeutically useful when one talks about generating different types of chimeric animals or of using newly developed genetic engineering tools (the CRISPR CAS9 system found in prokaryotes), that greatly simplify and improve the specificity of the targeted manipulation of specific genes (link). In these studies the animal of choice is not mice, but pigs – which because of their larger size produce organs for transplantion that are similar in size to the organs of people (see link). While similar in size, there are two issues that complicate pig to human organ transplantation: first there is the human immune system mediated rejection of foreign tissue and second there is the possibility that transplantation of porcine organs will lead to the infection of the human recipient with porcine retroviruses.

The issue of rejection (pig into human), always a serious problem, is further exacerbated by the presence in pigs of a gene encoding the enzyme α-1,3 galactosyl transferase (GGTA1). GGTA1 catalyzes the addition of the gal-epitope to a number of cell surface proteins. The gal-epitope is “expressed on the tissues of all mammals except humans and subhuman primates, which have antibodies against the epitope” (7). The result is that pig organs provoke an extremely strong immune (rejection) response in humans. The obvious technical fix to this (and related problems) is to remove the gal-epitope from pig cells by deleting the GGTA1 enzyme (see 8). It is worth noting that “organs from genetically engineered animals have enjoyed markedly improved survivals in non-human primates” (see Sachs & Gall, 2009).

pig to human The second obstacle to pig → human transplantation is the presence of retroviruses within the pig genome. All vertebrate genomes, including those of humans, contain many inserted retroviruses; almost 50% of the human genome is retrovirus-derived sequence (an example of unintelligent design if ever there was one). Most of these endogenous retroviruses are “under control” and are normally benign (see 9). The concern, however, is that the retroviruses present in pig cells could be activated when introduced into humans. To remove (or minimize) this possibility, Niu et al set out to use the CRISPR CAS9 system to delete these porcine endogenous retroviral sequences (PERVs) from the pig genome; they appear to have succeeded, generating a number of genetically modified pigs without PERVs (see 10). The hope is that organs generated from PERV-minus pigs from which antigen-generating genes, such as α-1,3 galactosyl transferase, have also been removed or inactivated together with more sophisticated inhibitors of tissue rejection, will lead to an essentially unlimited supply of pig organs that can be used for heart and other organ transplantation (see 11), and so alleviate the delays in transplantation, and so avoid deaths in sick people and the often brutal and criminal harvesting of organs carried out in some countries.

The final strategy being explored is to use genetically modified hosts and patient derived iPSCs to generate fully patient compatible human organs. To date, pilot studies have been carried out, apparently successfully, using rat embryos with mouse stem cells (see 12 and 13), with much more preliminary studies using pig embryos and human iPSCs (see 14). The approach involves what is known as chimeric embryos. In this case, host animals are genetically modified so that they cannot generate the organ of choice. Typically this is done by mutating a key gene that encodes a transcription factor directly involved in formation of the organ; embryos missing pancreas, kidney, heart, human pig embryo chimera or eyes can be generated. In an embryo that cannot make these organs, which can be a lethal defect, the introduction of stem cells from an animal that can form these organs can lead to the formation of an organ composed primarily of cells derived from the transplanted (human) cells.

At this point the strategy appears to work reasonably well for mouse-rat chimeras, which are much more closely related, evolutionarily, than are humans and pigs. Early studies on pig-human chimeras appear to be dramatically less efficient. At this point, Jun Wu has been reported as saying of human-pig chimeras that “we estimate [each had] about one in 100,000 human cells” (see 15), with the rest being pig cells. The bottom line appears to be that there are many technical hurdles to over-come before this method of developing patient-compatible human organs becomes feasible. Closer to reality are PERV-free/gal-antigen free pig-derived, human compatible organs. The reception of such life-saving organs by the general public, not to mention religious and philosophical groups that reject the consumption of animals in general, or pigs in particular, remains to be seen.

figures reinserted & minor edits 23 October 2020 – new link 17 December 2020.
references cited

Reverse Dunning-Kruger effects and science education

The Dunning-Kruger (DK) effect is the well-established phenomenon that people tend to over estimate their understanding of a particular topic or their skill at a particular task, often to a dramatic degree [link][link]. We see examples of the DK effect throughout society; the current administration (unfortunately) and the nutritional supplements / homeopathy section of Whole Foods spring to mind as examples. But there is a less well-recognized “reverse DK” effect, namely the tendency of instructors, and a range of other public communicators, to over-estimate what the people they are talking to are prepared to understand, appreciate, and accurately apply. The efforts of science communicators and instructors can be entertaining but the failure to recognize and address the reverse DK effect results in ineffective educational efforts. These efforts can themselves help generate the illusion of understanding in students and the broader public (discussed here). While a confused understanding of the intricacies of cosmology or particle physics can be relatively harmless in their social and personal implications, similar misunderstandings become personally and publicly significant when topics such as vaccination, alternative medical treatments, and climate change are in play.

There are two synergistic aspects to the reverse DK effect that directly impact science instruction: the need to understand what one’s audience does not understand together with the need to clearly articulate the conceptual underpinnings needed to understand the subject to be taught. This is in part because modern science has, at its core, become increasingly counter-intuitive over the last approximately 100 years or so, a situation that can cause serious confusions that educators must address directly and explicitly. The first reverse DK effect involves the extent to which the instructor (and by implication the course and textbook designer) has an accurate appreciation of what students think or think they know, what ideas they have previously been exposed to, and what they actually understand about the implications of those ideas. Are they prepared to learn a subject or does the instructor first have to acknowledge and address conceptual confusions and build or rebuild base concepts? While the best way to discover what students think is arguably a Socratic discussion, this only rarely occurs for a range of practical reasons. In its place, a number of concept inventory-type testing instruments have been generated to reveal whether various pre-identified common confusions exist in students’ thinking. Knowing the results of such assessments BEFORE instruction can help customize how the instructor structures the learning environment and content to be presented and whether the instructor gives students the space to work with these ideas to develop a more accurate and nuanced understanding of a topic. Of course, this implies that instructors have the flexibility to adjust the pace and focus of their classroom activities. Do they take the time needed to address student issues or do they feel pressured to plow through the prescribed course content, come hell, high water, or cascading student befuddlement.

A complementary aspect of the reverse DK effect, well-illustrated in the “why magnets attract” interview with the physicist Richard Feynman, is that the instructor, course designer, or textbook author(s) needs to have a deep and accurate appreciation of the underlying core knowledge necessary to understand the topic they are teaching. Such a robust conceptual understanding makes it possible to convey the complexities involved in a particular process and explicitly values appreciating a topic rather than memorizing it. It focuses on the general, rather than the idiosyncratic. A classic example from many an introductory biology course is the difference between expecting students to remember the steps in glycolysis or the Krebs cycle reaction system, as opposed to the general principles that underlie the non-equilibrium reaction networks involved in all biological functions, a reaction network based on coupled chemical reactions and governed by the behaviors of thermodynamically favorable and unfavorable reactions. Without a explicit discussion of these topics, all too often students are required to memorize names without understanding the underlying rationale driving the processes involved; that is, why the system behaves as it does. Instructors also give false “rubber band” analogies or heuristics to explain complex phenomena (see Feynman video 6:18 minutes in). A similar situation occurs when considering how molecules come to associate and dissociate from one another, for example in the process of regulating gene expression or repairing mutations in DNA. Most textbooks simply do not discuss the physiochemical processes involved in binding specificity, association, and dissociation rates, such as the energy changes associated with molecular interactions and thermal collisions (don’t believe me? look for yourself!). But these factors are essential for a student to understand the dynamics of gene expression [link], as well as the specificity of modern methods involved in genetic engineering, such as restriction enzymes, polymerase chain reaction, and CRISPR CAS9-mediated mutagenesis. By focusing on the underlying processes involved we can avoid their trivialization and enable students to apply basic principles to a broad range of situations. We can understand exactly why CRISPR CAS9-directed mutagenesis can be targeted to a single site within a multibillion-base pair genome.

Of course, as in the case of recognizing and responding to student misunderstandings and knowledge gaps, a thoughtful consideration of underlying processes takes course time, time that trades the development of a working understanding of core processes and principles for broader “coverage” of frequently disconnected facts, the memorization and regurgitation of which has been privileged over understanding why those facts are worth knowing. If our goal is for students to emerge from a course with an accurate understanding of the basic processes involved rather than a superficial familiarity with a plethora of unrelated facts, however, a Socratic interaction with the topic is essential. What assumptions are being made, where do they come from, how do they constrain the system, and what are their implications? Do we understand why the system behaves the way it does? In this light, it is a serious educational mystery that many molecular biology / biochemistry curricula fail to introduce students to the range of selective and non-selective evolutionary mechanisms (including social and sexual selection – see link), that is, the processes that have shaped modern organisms.

Both aspects of the reverse DK effect impact educational outcomes. Overcoming the reverse DK effect depends on educational institutions committing to effective and engaging course design, measured in terms of retention, time to degree, and a robust inquiry into actual student learning. Such an institutional dedication to effective course design and delivery is necessary to empower instructors and course designers. These individuals bring a deep understanding of the topics taught and their conceptual foundations and historic development to their students AND must have the flexibility and authority to alter the pace (and design) of a course or a curriculum when they discover that their students lack the pre-existing expertise necessary for learning or that the course materials (textbooks) do not present or emphasize necessary ideas. Radiation-kills-in-Boulder Unfortunately, all too often instructors, particularly in introductory level college science courses, are not the masters of their ships; that is, they are not rewarded for generating more effective course materials. An emphasis on course “coverage” over learning, whether through peer-pressure, institutional apathy, or both, generates unnecessary obstacles to both student engagement and content mastery. To reverse the effects of the reverse DK effect, we need to encourage instructors, course designers, and departments to see the presentation of core disciplinary observations and concepts as the intellectually challenging and valuable endeavor that it is. In its absence, there are serious (and growing) pressures to trivialize or obscure the educational experience – leading to the socially- and personally-damaging growth of fake knowledge.

empty images holders removed, new image added – 17 December 2020

Is it time to start worrying about conscious human “mini-brains”?

A human iPSC cerebral organoid in which pigmented retinal epithelial cells can be seen (from the work of McClure-Begley et al.). Also see “Can lab-grown brains become conscious?” by Sara Readon Nature 2020.

The fact that experiments on people are severely constrained is a major obstacle in understanding human development and disease. Some of these constraints are moral and ethical and clearly appropriate and necessary given the depressing history of medical atrocities. Others are technical, associated with the slow pace of human development. The combination of moral and technical factors has driven experimental biologists to explore the behavior of a wide range of “model systems” from bacteria, yeasts, fruit flies, and worms to fish, frogs, birds, rodents, and primates. Justified by the deep evolutionary continuity between these organisms (after all, all organisms appear to be descended from a single common ancestor and share many molecular features), experimental evolution-based studies of model systems have led to many therapeutically valuable insights in humans – something that I suspect a devotee of intelligent design creationism would be hard pressed to predict or explain (post link).

While humans are closely related to other mammals, it is immediately obvious that there are important differences – after all people are instantly recognizable from members of other closely related species and certainly look and behave differently from mice. For example, the surface layer of our brains is extensively folded (they are known as gyrencephalic) while the brain of a mouse is smooth as a baby’s bottom (and referred to as lissencephalic). In humans, the failure of the brain cortex to fold is known as lissencephaly, a disorder associated with severe neurological defects. With the advent of more and more genomic sequence data, we can identify human specific molecular (genomic) differences. Many of these sequence differences occur in regions of our DNA that regulate when and where specific genes are expressed. Sholtis & Noonan (1) provide an example: the HACNS1 locus is a 81 basepair region that is highly conserved in various vertebrates from birds to chimpanzees; there are 13 human specific changes in this sequence that appear to alter its activity, leading to human-specific changes in the expression of nearby genes (↓). At this point ~1000 genetic elements that are different in humans compared to other vertebrates have been identified and more are likely to emerge (2). Such human-specific changes can make modeling human-specific behaviors, at the cellular, tissue, organ, and organism level, in non-human model systems difficult and problematic (3, 4). It is for this reason that scientists have attempted to generate better human specific systems.

human sequence divergence

One particularly promising approach is based on what are known as embryonic stem cells (ESCs) or pluripotent stem cells (PSCs). Human embryonic stem cells are generated from the inner cell mass of a human embryo and so involve the destruction of that embryo – which raises a number of ethical and religious concerns as to when “life begins” (5). Human pluripotent stem cells are isolated from adult tissues but in most cases require invasive harvesting methods that limit their usefulness. Both ESCs and PSCs can be grown in the laboratory and can be induced to differentiate into what are known as gastruloids. Such gastruloids can develop anterior-posterior (head-tail), dorsal-ventral (back-belly), and left-right axes analogous to those found in embryos (6) and adults (top panel ↓). In the case of PSCs, the gastruloid (bottom panel ↓) is essentially a twin of the organism from which the PSCs were derived, a situation that raises difficult questions: is it a distinct individual, is it the property of the donor or the creation of a technician. The situation will be further complicated if (or rather, when) it becomes possible to generate viable embryos from such gastruloids.

Axes

gastruloid-embryo-comparison The Nobel prize winning work of Kazutoshi Takahashi and Shinya Yamanaka (7), who devised methods to take differentiated (somatic) human cells and reprogram them into ESC/PSC-like cells, cells known as induced pluripotent stem cells (iPSCs)(8), represented a technical breakthrough that jump-started this field. While the original methods derived sample cells from tissue biopsies, it is possible to reprogram kidney epithelial cells recovered from urine, a non-invasive approach (9, 10). Subsequently, Madeline Lancaster, Jurgen Knōblich, and colleagues devised an approach by which such cells could be induced to form what they termed “cerebral organoids” (although Yoshiki Sasai and colleagues were the first to generate neuronal organoids); they used this method to examine the developmental defects associated with microencephaly (11). The value of the approach was rapidly recognized and a number of studies on human conditions, including lissencephaly (12), Zika-virus infection-induced microencephaly (13), and Down’s syndrome (14); investigators have begun to exploit these methods to study a range of human diseases – and rapid technological progress is being made.

The production of cerebral organoids from reprogrammed human somatic cells has also attracted the attention of the media (15). While “mini-brain” is certainly a catchier name, it is a less accurate description of a cerebral organoid, itself possibly a bit of an overstatement, since it is not clear exactly how “cerebral” such organoids are. For example, the developing brain is patterned by embryonic signals that establish its asymmetries; it forms at the anterior end of the neural tube (the nascent central nervous system and spinal cord) and with distinctive anterior-posterior, dorsal-ventral, and left-right asymmetries, something that simple cerebral organoids do not display. Moreover, current methods for generating cerebral organoids involve primarily what are known as neuroectodermal cells – our nervous system (and that of other vertebrates) is a specialized form of the embryo’s surface layer that gets internalized during development. In the embryo, the developing neuroectoderm interacts with cells of the circulatory system (capillaries, veins, and arteries), formed by endothelial cells and what are known as pericytes that surround them. These cells, together with interactions with glial cells (astrocytes, a non-neuronal cell type) combine to form the blood brain barrier. Other glial cells (oligodendrocytes) are also present; in contrast, both types of glia (astrocytes and oligodendrocytes) are rare in the current generation of cerebral organoids. Finally, there are microglial cells, immune system cells that originate from outside the neuroectoderm; they invade and interact with neurons and glia as part of the brain’s dynamic neural capillary and neurons system. The left panel of the figure shows, in highly schematic form how these cells interact (16). The right panel is a drawing of neural tissue stained by the Golgi method (17), which reveals ~3-5% of the neurons present. There are at least as many glial cells present, as well as microglia, none of which are visible in the image. At this point, cerebral organoids typically contain few astrocytes and oligodendrocytes, no vasculature, and no microglia. Moreover, they grow to be about 1 to 3 mm in diameter over the course of 6 to 9 months; that is significantly smaller in volume than a fetal or newborn’s brain. While cerebral organoids can generate structures characteristic of retinal pigment epithelia (top figure) and photo-responsive neurons (18), such as those associated with the retina, an extension of the brain, it is not at all clear that there is any significant sensory input into the neuronal networks that are formed within a cerebral organoid, or any significant outputs, at least compared to the role that the human brain plays in controlling bodily and mental functions.

The reasonable question, then, must be whether a cerebral organoid, which is a relatively simple system of cells (although itself complex), is conscious. It becomes more reasonable as increasingly complex systems are developed, and such work is proceeding apace. Already researchers are manipulating the developing organoid’s environment to facilitate axis formation, and one can anticipate the introduction of vasculature. Indeed, the generation of microglia-like cells from iPSCs has been reported; such cells can be incorporated into cerebral organoids where they appear to respond to neuronal damage in much the same way as microglia behave in intact neural tissue (19).

We can ask ourselves, what would convince us that a cerebral organoid, living within a laboratory incubator, was conscious? How would such consciousness manifest itself? Through some specific pattern of neural activity, perhaps? As a biologist, albeit one primarily interested in molecular and cellular systems, I discount the idea, proposed by some physicists and philosophers as well as the more mystical, that consciousness is a universal property of matter (20,21). I take consciousness to be an emergent property of complex neural systems, generated by evolutionary mechanisms, built during embryonic and subsequent development, and influenced by social interactions (BLOG LINK) using information encoded within the human genome (something similar to this: A New Theory Explains How Consciousness Evolved). While a future concern, in a world full of more immediate and pressing issues, it will be interesting to listen to the academic, social, and political debate on what to do with mini-brains as they grow in complexity and perhaps inevitably, towards consciousness.

Footnotes and references

Thanks to Rebecca Klymkowsky, Esq. and Joshua Sanes, Ph.D. for editing and disciplinary support. Minor updates and the reintroduction of figures 22 Oct. 2020.