Laboratory earthquake forecasting: A machine learning competition

Laboratory earthquake forecasting: A machine learning competition. Paul A. Johnson et al. Proceedings of the National Academy of Sciences, February 2, 2021 118 (5) e2011362118; https://doi.org/10.1073/pnas.2011362118

Abstract: Earthquake prediction, the long-sought holy grail of earthquake science, continues to confound Earth scientists. Could we make advances by crowdsourcing, drawing from the vast knowledge and creativity of the machine learning (ML) community? We used Google’s ML competition platform, Kaggle, to engage the worldwide ML community with a competition to develop and improve data analysis approaches on a forecasting problem that uses laboratory earthquake data. The competitors were tasked with predicting the time remaining before the next earthquake of successive laboratory quake events, based on only a small portion of the laboratory seismic data. The more than 4,500 participating teams created and shared more than 400 computer programs in openly accessible notebooks. Complementing the now well-known features of seismic data that map to fault criticality in the laboratory, the winning teams employed unexpected strategies based on rescaling failure times as a fraction of the seismic cycle and comparing input distribution of training and testing data. In addition to yielding scientific insights into fault processes in the laboratory and their relation with the evolution of the statistical properties of the associated seismic data, the competition serves as a pedagogical tool for teaching ML in geophysics. The approach may provide a model for other competitions in geosciences or other domains of study to help engage the ML community on problems of significance.

Keywords: machine learning competitionlaboratory earthquakesearthquake predictionphysics of faulting

What Did We Learn from the Kaggle Competition?

Previous work on seismic data from Earth (3) suggests that the underlying physics may scale from a laboratory fault to large fault systems in Earth. If this is indeed the case, improvements in our ability to predict earthquakes in the laboratory could lead to significant progress in time-dependent earthquake hazard characterization. The ultimate goal of the earthquake prediction challenge was to identify promising ML approaches for seismic data analysis that may enable improved estimates of fault failure in the Earth. In the following, we will discuss shortcomings of the competition but also key innovations that improved laboratory quake predictions and may be transposed to Earth studies.

The approaches employed by the winning teams included several innovations considerably different from our initial work on laboratory quake prediction (1). Team Zoo added synthetic noise to the input seismic data before feature computing and model training, thus making their models more robust to noise and more likely to generalize.

Team Zoo, JunKoda, and GloryorDeath only considered features that exhibited similar distributions between the training and testing data, thereby ensuring that nonstationary features could not be used in the learning phase and again, improving model generalization. We note that employing the distribution of the testing set input is a form of data snooping that effectively made the test set actually a validation set. However, the idea of employing only features with distributions that do not evolve over time is insightful and could be used for scientific purposes by comparing feature distribution between portions of training data, for example.

Perhaps most interestingly from a physical standpoint, the fifth team, Team Reza, changed the target to be predicted and endeavored to predict the seismic cycle fraction remaining instead of time remaining before failure. Because they did not employ the approach of comparing input distribution between training and testing sets as done by the first, second, and fourth teams, the performance impact from the prediction of normalized time to failure (seismic cycle fraction) was significant.

As in any level of statistics, more data are in general better and can improve model performance. Thus, had the competitors been given more training data, in principle scores may have improved. At the same time, there is an element of nonstationarity in the experiment because the fault gouge layer thins as the experiment progresses, and therefore, even an extremely large dataset would not lead to a perfect prediction. In addition, Kaggle keeps the public/private test set split in such a way as to not reward overfitting. No matter how large the dataset is, if a model iterates enough times on that dataset, it will not translate well into “the real world,” so the competition structure was designed to prevent that opportunity.

It is worth noting that the ML metric should be carefully considered. In Earth, it will be important to accurately predict the next quake as it approaches, but MAE treats each time step equally with respect to the absolute error making this challenging.

Individuals participate on the Kaggle platform for many reasons; the most common are the ability to participate in interesting and challenging projects in many different domains, the ability to learn and practice ML and data science skills, the ability to interact with others who are seeking the same, and of course, cash prizes. The astounding intellectual diversity the Kaggle platform attracted for this competition, with team representations from cartoon publishers, insurance agents, and hotel managers, is especially notable. In fact, none of the competition winners came from geophysics. Teams exhibit collective interaction, evidenced by the step changes in the MAE through time (Fig. 6), likely precipitated by communication through the discussion board and shared code.

The competition contributed to an accelerating increase in ML applications in the geosciences, has become an introductory problem for the geoscience community to learn different ML approaches, and is used for ML classes in geoscience departments. Students and researchers have used the top five approaches to compare the nuances of competing ML methods, as well as to try to adapt and improve the approaches for other applications.

Cats show no avoidance of people who behave negatively to their owner, unlike dogs

Chijiiwa, H., Takagi, S., Arahori, M., Anderson, J. R., Fujita, K., & Kuroshima, H. (2021). Cats (Felis catus) show no avoidance of people who behave negatively to their owner. Animal Behavior and Cognition, 8(1), 23-35. https://doi.org/10.26451/abc.08.01.03.2021

Rolf Degen's take: Unlike dogs, cats show no avoidance of people who behave negatively toward their owner

Abstract: Humans evaluate others based on interactions between third parties, even when those interactions are of no direct relevance to the observer. Such social evaluation is not limited to humans. We previously showed that dogs avoided a person who behaved negatively to their owner (Chijiiwa et al., 2015). Here, we explored whether domestic cats, another common companion animal, similarly evaluate humans based on third-party interactions. We used the same procedure that we used with dogs: cats watched as their owner first tried unsuccessfully to open a transparent container to take out an object, and then requested help from a person sitting nearby. In the Helper condition, this second person (helper) helped the owner to open the container, whereas in the Non-Helper condition the actor refused to help, turning away instead. A third, passive (neutral) person sat on the other side of the owner in both conditions. After the interaction, the actor and the neutral person each offered a piece of food to the cat, and we recorded which person the cat took food from. Cats completed four trials and showed neither a preference for the helper nor avoidance of the non-helper. We consider that cats might not possess the same social evaluation abilities as dogs, at least in this situation, because unlike the latter, they have not been selected to cooperate with humans. However, further work on cats’ social evaluation capacities needs to consider ecological validity, notably with regard to the species’ sociality.

Keywords: Cats, Social evaluation, Third-party interaction, Social cognition, Cat-human relationship, Domesticated animals

The stability of psychological adjustment among donor-conceived offspring in the U.S. National Longitudinal Lesbian Family Study from childhood to adulthood: Good adjustment in the long term

The stability of psychological adjustment among donor-conceived offspring in the U.S. National Longitudinal Lesbian Family Study from childhood to adulthood: differences by donor type. Nicola Carone et al. Fertility and Sterility, February 2 2021. https://doi.org/10.1016/j.fertnstert.2020.12.012

Rolf Degen's take: Having been conceived by an anonymous sperm donor did not interfere with identity development in children from lesbian parents

Abstract

Objective: To study differences by sperm donor type in the psychological adjustment of the U.S. National Longitudinal Lesbian Family Study (NLLFS) offspring across three time periods from childhood to adulthood.

Design: U.S.-based prospective cohort study.

Setting: Paper-and-pencil questionnaires and protected online surveys.

Patients: A cohort of 74 offspring conceived by lesbian parents using an anonymous (n = 26), a known (n = 26), or an open-identity (n = 22) sperm donor. Data were reported when offspring were ages 10 (wave 4), 17 (wave 5), and 25 (wave 6).

Main Outcome Measure: Achenbach Child Behavior Checklist administered to lesbian parents when offspring were ages 10 and 17 and the Achenbach Adult Self-Report administered to offspring at age 25.

Results: In both relative and absolute stability, no differences were found in internalizing, externalizing, and total problem behaviors by donor type over 15 years. However, both externalizing and total problem behaviors significantly declined from age 10 to 17 and then increased from age 17 to 25. Irrespective of donor type, among the 74 offspring, the large majority scored continuously within the normal range on internalizing (n = 62, 83.8%), externalizing (n = 62, 83.8%), and total problem behaviors (n = 60, 81.1%).

Conclusions: The results reassure prospective lesbian parents and provide policy makers and reproductive medicine practitioners with empirical evidence that psychological adjustment in offspring raised by lesbian parents is unrelated to donor type in the long term.

Keywords: Sperm donationanonymityopen-identitypsychological adjustmentlesbian parents

From 2019... Simple hair‐like feathers served as insulating pelage, but the first feathers with complex branching structures and a plainer form evolved for the purpose of sexual display

From 2019... Feather evolution exemplifies sexually selected bridges across the adaptive landscape. W. Scott Persons  Philip J. Currie. Evolution, July 19 2019. https://doi.org/10.1111/evo.13795

h/t David Schmitt  @PsychoSchmitt

Abstract: Over the last two decades, paleontologists have pieced together the early evolutionary history of feathers. Simple hair‐like feathers served as insulating pelage, but the first feathers with complex branching structures and a plainer form evolved for the purpose of sexual display. The evolution of these complex display feathers was essential to the later evolution of flight. Feathers illustrate how sexual selection can generate complex novel phenotypes, which are then available for natural selection to modify and direct toward novel functions. In the longstanding metaphor of the adaptive landscape, sexual selection is a means by which lineages resting on one adaptive peak may gradually bridge a gap to another peak, without the landscape itself being first altered by environmental changes.

Individuals with depression express more distorted thinking on social media

Individuals with depression express more distorted thinking on social media. Krishna C. Bathina, Marijn ten Thij, Lorenzo Lorenzo-Luaces, Lauren A. Rutter & Johan Bollen. Nature Human Behaviour, February 11 2021. https://www.nature.com/articles/s41562-021-01050-7

Abstract: Depression is a leading cause of disability worldwide, but is often underdiagnosed and undertreated. Cognitive behavioural therapy holds that individuals with depression exhibit distorted modes of thinking, that is, cognitive distortions, that can negatively affect their emotions and motivation. Here, we show that the language of individuals with a self-reported diagnosis of depression on social media is characterized by higher levels of distorted thinking compared with a random sample. This effect is specific to the distorted nature of the expression and cannot be explained by the presence of specific topics, sentiment or first-person pronouns. This study identifies online language patterns that are indicative of depression-related distorted thinking. We caution that any future applications of this research should carefully consider ethical and data privacy issues.

Discussion

In a sample of online individuals, we used a theory-driven approach to measure the prevalence of linguistic markers that may indicate cognitive vulnerability to depression, according to CBT theory. We defined a set of CDS that we grouped along 12 widely accepted types of distorted thinking and compared their prevalence between two cohorts of Twitter users—the first included individuals who reported that they received a clinical diagnosis of depression and the second was a similar random sample.

As hypothesized, the individuals in the D cohort use significantly more CDS in their online language compared with individuals in the R cohort, particularly schemata associated with ‘personalizing’ and ‘emotional reasoning’. We observed significantly increased levels of CDS across nearly all cognitive distortion types, sometimes more than twice as much, but did not find a statistically significant increase in prevalence among the D cohort for two specific types, namely ‘fortune-telling’ and ‘catastrophizing’. This may be due to the difficulty of capturing these specific cognitive distortions in the form of a set of 1–5-grams—their expression in language can involve an interactive process of conversation and interpretation. Notably, our findings are not explained by the use of FPPs or more negatively loaded language. These results shed a light on the degree to which depression-related language of cognitive distortions are manifested in the colloquial language of social media platforms. This is of social relevance given that these platforms are specifically designed to propagate information through the social ties that connect individuals on a global scale.

An advantage of studying theory-driven differences between the language of individuals with and without depression, in contrast to a purely data-driven or machine learning approach, is that we can explicitly use the principles underpinning CBT to understand the cognitive and lexical components that may shape depression. Cognitive behavioural therapists have developed a set of strategies to challenge the distorted thinking patterns that are characteristic of depression. Preliminary findings suggest that specific language can be related to specific therapeutic practices and seems to be related to outcomes48. However, these practices have been largely shaped by a clinical understanding and not necessarily informed by objective measures of how patterns of language reflect cognitive distortions, which could be harnessed to facilitate the path of recovery.

Our results suggest a path for mitigation and intervention, including applications that engage individuals with mood disorders, such as major depressive disorder, through social media platforms and that challenge particular expressions and types of depression-related language. Future characterization of the relationship between depression-related language and mood may help in the development of automated interventions (such as ‘chatbots’) or suggest promising targets for psychotherapy. Another approach that has shown promise in leveraging social media for the treatment of mental health problems involves crowdsourcing the responses to cognitively distorted content49. These types of applications have the potential to be more-scalable mental health interventions compared with existing approaches such as face-to-face psychotherapy50. The extent to which user CDS prevalence can be used as a passive index of vulnerability to depression that may be expected to change with treatment could also be explored. Insofar as online language can be considered to be an index of cognitive vulnerability to depression, a better understanding of online language may help to tailor treatments, especially internet-based treatments, to the more-specific needs of individuals. For example, interventions that target depression-related thinking and language may be well-suited for individuals with depression who express relatively higher levels of these distortions, whereas interventions that target other mechanisms (such as physical activity, circadian rhythm) may be better suited for individuals who do not show relatively higher levels of CDS. More research towards understanding differences in language patterns in depression and related disorders, such as anxiety disorders, is recommended. However, when implementing these types of approaches, ethical considerations and privacy issues have to be adequately addressed38,39.

Several limitations of our theory-driven approach should be considered. First, we relied on individuals reporting their personal clinical depression diagnoses on social media. Although we verified that the statement indeed pertains to a clinical diagnosis, we do not have verification of the diagnosis itself nor of its accuracy. This may introduce individuals into the D cohort who might not have been diagnosed with depression or accurately diagnosed. Vice versa, we have no verification that individuals in our random sample do not suffer from depression. However, the potential inaccuracy of this inclusion criterion will probably reduce the difference in depression rates between the two cohorts and, therefore, reduce the observed effect sizes (PR values between cohorts) due to the larger heterogeneity of our sample. As a consequence, our results are probably not an artefact of the accuracy of our inclusion criterion. Second, our approach is limited to discovering only individuals who are willing to disclose their diagnosis on social media. As this might skew our D cohort to a subgroup of individuals suffering from depression, we recommend caution when generalizing our findings to the level of all individuals who have depression. Third, our lexicon of CDS was composed and approved by a panel of ten experts who may have been only partially successful in capturing all of the n-grams used to express distorted ways of thinking. On a related note, the use of CDS n-grams implies that we measure distorted thinking by proxy, namely through language, and our observations may be therefore be affected by linguistic and cultural factors. Common idiosyncratic or idiomatic expressions may syntactically represent a distorted form of thinking, but no longer do so in practice. For example, an expression such as ‘literally the worst’ may be commonly used to express dismay, without necessarily involving the speaker experiencing a distorted mode of thinking. Thus, the presence of a CDS does not point to a cognitive distortion per se. Fourth, both cohorts were sampled from Twitter, one of the leading social media platforms, the use of which may be associated with higher levels of psychopathology and reduced well-being51,52,53. We may therefore be observing increased or biased rates of distorted thinking in both cohorts as a result of platform effects. However, we report relative prevalence numbers with respect to a carefully construed random sample also taken from Twitter, which probably compensates for this effect and the effect that individuals with depression might be more active than their random counterparts. Furthermore, recent analysis indicates that representative samples with respect to psychological phenomena can be obtained from social media content54. This is an important discussion in computational social science that will continue to be investigated. Data-driven approaches that analyse natural language in real-time will continue to complement theory-driven work such as ours.

As we analysed individuals on the basis of inferred health-related information, we want to stress some additional considerations regarding ethical research practices and data privacy30,38,39. We limited our investigation strictly to comparing, in the aggregate, the publicly shared language of two deidentified cohorts of individuals (individuals who report that they have been diagnosed with depression and a random sample). We carefully deidentified all obtained data to protect user privacy and performed our analysis under the constraints of two IRB protocols (IU IRB Protocols 2010371843 and 1707249405). Whereas the outcomes of our analysis could contribute to a better understanding of depression as a mental health disorder, they could also inform approaches that detect traces of mental health issues in the online language of individuals, and as such contribute to future detection, diagnostics and intervention efforts. This may raise important ethical and user privacy concerns as well as risk of harm, including but not limited to the right to privacy, data ownership and transparency. For example, even though social media data are technically public, individuals do not necessarily realize nor consent to particular retrospective analysis when they share information on their public accounts55 nor can they consent to how these data may be leveraged in future approaches that may involve individualized interactions and inventions. Considering existing evidence that individuals are more willing to share biomedical data than social media data56, in future research, we hope to reach a larger sample of individuals who understand public data availability and increase transparency through a carefully managed consent process. We acknowledge that these considerations are part of an active and ongoing discussion in our community that we encourage and that we hope our research may contribute to.

We emphasize that not all use of CDS n-grams reflects depressive thinking, as these phrases are part of normal English usage, and it would therefore be wrong to try to diagnose depression merely on the basis of use of one or more such phrases. Such an approach would, as well as being inaccurate, potentially lead to harm in terms of stigmatizing individuals.

Our results showed that the exposure to fear chemosignals (vs. rest chemosignals and a no-sweat condition) while not changing vigilance behavior leads to faster answers to threatening events

The Function of Fear Chemosignals: Preparing for Danger. Nuno Gomes, Gün R Semin. Chemical Senses, bjab005, February 11 2021. https://doi.org/10.1093/chemse/bjab005

Abstract: It has been shown that the presence of conspecifics modulates human’s vigilance strategies as is the case with animal species. Mere presence has been found to reduce vigilance. However, animal research has also shown that chemosignals (e.g., sweat) produced during fear-inducing situations modulates individuals’ threat detection strategies. In the case of humans, little is known about how exposure to conspecifics’ fear chemosignals modulates vigilance and threat detection effectiveness. The present study (N= 59) examined how human fear chemosignals affect vigilance strategies and threat avoidance in its receivers. We relied on a paradigm that simulates a “foraging under threat” situation in the lab, integrated with an eye-tracker to examine the attention allocation. Our results showed that the exposure to fear chemosignals (vs. rest chemosignals and a no-sweat condition) while not changing vigilance behavior leads to faster answers to threatening events. In conclusion, fear chemosignals seem to constitute an important warning signal for human beings, possibly leading its receiver to a readiness state that allows faster reactions to threat-related events.

Keywords: Vigilance, Fear chemosignals, Olfaction, Threat detection, Eye-tracking

Reduced decision bias and more rational decision making following cortex damage

Reduced decision bias and more rational decision making following ventromedial prefrontal cortex damage. Sanjay Manohar et al. Cortex, February 11 2021. https://doi.org/10.1016/j.cortex.2021.01.015

Summary: Human decisions are susceptible to biases, but establishing causal roles of brain areas has proved to be difficult. Here we studied decision biases in 17 people with unilateral medial prefrontal cortex damage and a rare patient with bilateral ventromedial prefrontal cortex (vmPFC) lesions. Participants learned to choose which of two options was most likely to win, and then bet money on the outcome. Thus, good performance required not only selecting the best option, but also the amount to bet. Healthy people were biased by their previous bet, as well as by the unchosen option’s value. Unilateral medial prefrontal lesions reduced these biases, leading to more rational decisions. Bilateral vmPFC lesions resulted in more strategic betting, again with less bias from the previous trial, paradoxically improving performance overall. Together, the results suggest that vmPFC normally imposes contextual biases, which in healthy people may actually be suboptimal in some situations.

Discussion

In this study we used a novel reversal learning task in which participants made post-decision wagers on their choices, thereby providing a measure of their confidence in winning, and also rated their surprise at outcomes (Fig.1). Analysis was performed on both performance data as well as with a computational model of value learning. In healthy volunteers, bets tracked the expected chance of winning (Fig.4A), but also showed strong biases: People’s bets tended to be similar to their bets on the previous trial, and were higher when the unchosen option was less likely to win. Patients with unilateral mPFC lesions bet more overall (Fig.3B), but showed weaker biases from the previous trial and from the unchosen option. The bilateral patient MJ also showed a weaker bias from the previous trial (Fig.4B), but crucially had a stronger effect of the chosen option’s probability of winning (Fig.4A). This meant that he won more than any other healthy volunteer or unilateral patient on this task (Fig.2A), despite no difference in learning which option was better. Thus, his performance can be seen as exhibiting a more rational betting strategy than in healthy people.

A large body of evidence has revealed that many aspects of human decision making are seemingly irrational, driven by biases that appear to lead to suboptimal outcomes (Tversky and Kahneman, 1974; De Martino et al., 2006; Talluri et al., 2018; Urai et al., 2019). Evidence that some of these biases are driven by normal cognitive operations underpinned by specific brain processes (De Martino et al., 2006; Wimmer and Shohamy, 2012) raises the possibility that damage to the brain might paradoxically reduce such biases (Akrami et al., 2018; Kapur, 1996) and perhaps lead to more rational behaviour. However, to date only limited causal evidence for such a possibility exists in humans (Greene, 2007; Knoch et al., 2006; Koenigs et al., 2007).

The findings presented here show that it is indeed possible for more rational decision making to emerge – at least on a value based reversal learning task – after bilateral vmPFC lesions. This is not to say that all decisions and behaviours become more rational after such brain damage. Clearly, although he managed to continue to work in a demanding job, patient MJ showed evidence of dysfunction in social cognition and some aspects of decision making and judgment in everyday life, just as previous reported cases (Eslinger and Damasio, 1985; Bechara et al., 2000; Berlin et al., 2004; Shamay-Tsoory et al., 2005).

There is some previous circumstantial evidence that mPFC lesions may reduce decision biases. For example, patients with mPFC damage show smaller biases in probabilistic estimation (O’Callaghan et al., 2018), reduced affective contributions to reasoning (Shamay-Tsoory et al., 2005), and may indeed make more utilitarian moral judgements, suggesting more rational valuation with less affective bias (Ciaramelli et al., 2007; Koenigs et al., 2007; Krajbich et al., 2009). These effects might be underpinned by a more general increase in rationality after damage to this region. One possible explanation for this is that individuals with vmPFC lesions might be free of affective biases that normally contribute to such decision making but this remains to be established.

In line with this, Shiv et al. (2005) asked patients with a variety of lesions (amygdala, orbitofrontal and insula) to opt in or out of gambles with positive expected value. Controls tended to opt out especially after a loss, whereas the patients continued to bet, thus winning more. This can be compared to our win-stay analysis (Fig.3F), where MJ bet more than controls on win-stay choices, but did not bet less on lose-switch choices. Further evidence that biases can depend on specific brain areas comes from patients with insula damage, who may lose the normal tendency towards the gamblers’ fallacy (Clark et al., 2014). With this bias, participants tend to re-choose an option that previously lost (because the history of wins should balance out on average). Transcranial stimulation to lateral prefrontal cortex increases this bias (Xue et al., 2012). In our study, there is a possible analogy with the unchosen option effect (Fig.4C), where people bet less when the alternative was valuable (perhaps also because the two options should balance out on average). Unilateral ventromedial patients lost this bias. However, in our task, lesions did not affect the option decisions themselves.

Biases from previous trials may rely on information retained in working memory. Thus an important null result is that the bilateral patient was unimpaired in working memory accuracy (Table S2). He had considerable difficulty remembering verbal lists (Table 1). This memory deficit might have contributed to his lack of trial history biases. However, against this possibility, performance was normal on a specific working memory task, suggesting that the previous bet effect was not simply memory-related. Furthermore, his normal learning and decision-making indicate he was integrating and retaining the specific value information involved in the biases, making memory deficits less likely to contribute. One interpretation of the loss of bias could be that medial frontal areas are required for normal integration of the biasing or interfering information into the current decision. An alternative interpretation is that normal biases are driven by suboptimal heuristics, and that medial frontal lesions abolish these heuristics.

[Table 1. Summary of standardised neuropsychological scores for patient MJ. Impairments were seen in the verbal learning task for short delay recall and yes-no recognition. WAIS: Wechsler adult intelligence scale; WMS: Wechsler memory scale; WM: working memory; CVLT: California verbal learning test; DKEFS: Delis-Kaplan executive function system; GNT: Warrington graded naming test. Red indicates scores in the “extremely low” range (<2nd centile), and pink indicates scores in the borderline range (<10th centile).]

Patients with vmPFC/OFC lesions have previously been shown to bet more under uncertainty (Clark et al., 2008), being generally less risk averse (Bechara et al., 2000; Levens et al., 2014), and our results directly support this finding. However, bets reflect a combination of general risk seeking, confidence, biases and strategic factors. In our study, increased betting alone was insufficient to explain the bilateral patient’s advantage in this task. Instead, reduced biases may have permitted strategic betting, such as the hot hand effect or loss chasing. Interestingly, a previous study had identified that dorsomedial prefrontal lesions can increase the bias caused by eye movements during decisions (Vaidya and Fellows, 2015), but to our knowledge, no human studies have shown reduced biases after lesions in the way demonstrated here.

Information about unchosen options and recent actions may be disrupted by medial or orbitofrontal lesions (Buckley et al., 2009; Levens et al., 2014), which might thus account for the reduced biases in unilateral patients. The effect parallels a recent rodent study where parietal inactivation also paradoxically improved performance, by reducing the active bias from previous trials (Akrami et al., 2018). However, it is unclear why bilateral lesions did not attenuate this bias in MJ. Our finding of larger surprise differences between winning and losing may also match previous reports of increased emotional responses to stochastic outcomes after vmPFC lesions (Levens et al., 2014) and could parallel increases in reward sensitivity observed in these patients (Manohar and Husain, 2016).

Intriguingly, we found no consistent effects on option-selection in this task. Previous studies of classical reversal learning in patients with vmPFC lesions have shown varied effects. Patients tend to perseverate, maladaptively repeating their previous choices even after reward contingencies reverse (Fellows and Farah, 2003; Rolls et al., 1994), but other studies have found only a marginal effect (Daum et al., 1991), and yet others showed normal performance after unilateral lesions, but impaired reversal after bilateral lesions (Hornak et al., 2004). This is consistent with detailed studies in animals suggesting that impairments after OFC lesions may be mild, with medial lesions only impairing performance when discrimination is harder (Izquierdo et al., 2017; Rudebeck and Murray, 2011), and impairments potentially improved by further lesions (Stalnaker et al., 2007). Yet other work has demonstrated that vmPFC lesions produce unstable choices while preserving subjective valuation of single objects (Henri-Bhargava et al., 2012). However, we did not find any deficits in value-based selection of options in our task. This could be because the paradigm used here crucially tests the use of learned values, rather than subjective valuation or rule-following.

In non-human primate studies, brain areas encoding the values of options also encode decision confidence, such as OFC (Kepecs et al., 2008). In humans, fMRI activation increases with decision confidence in vmPFC (De Martino et al., 2013; Lebreton et al., 2015; Rolls et al., 2010; Yokoyama et al., 2010). Although some studies have demonstrated inaccurate confidence judgements after prefrontal lesions (Fleming et al., 2014), others find no deficits even with bilateral lesions (Lemaitre et al., 2018). Remarkably, disrupting anterior PFC with TMS can actually improve metacognitive confidence judgements (Shekhar and Rahnev, 2018). Thus, if medial PFC encodes variables that might bias valuation, lesions to this area should paradoxically improve performance in some situations, as observed here.

Of course, human lesion studies are inherently limited by the possibility of damage not visible on the MRI scans. Although all patients reported here had brain haemorrhages affecting the mPFC, with very little damage outside this region, the bilateral patient had suffered a traumatic injury, followed by haemorrhage. It is possible that this resulted in a different pattern of microscopic damage: although traumatic injuries may appear focal, often the functional damage can be quite widespread. This limits the conclusions that can be drawn about the causal role of medial frontal cortex specifically. However we suggest that the most likely explanation is the bilateral nature of his lesions: reward value is usually considered to be represented bilaterally in OFC (Hampton and O’Doherty, 2007; Rolls, 2015), suggesting that unilateral lesions are less likely to show manifest impairments. One difficulty with interpreting lesion studies is whether the changes reflect direct lesion effects, or compensatory strategies. The chronic nature of his lesion may be a key difference between MJ and other studies demonstrating deficits in reversal learning after vmPFC lesions (Fellows and Farah, 2003; Rolls et al., 1994). This may have allowed recovery and adaptation, leading to his strategic betting pattern. In this case, it is unclear whether it is vmPFC loss per se, or the network-level consequences of this, that attenuates biases. Functional imaging studies in patients might potentially shed light on this in the future.

In summary, the results suggest that vmPFC may drive biases in healthy people. A patient with bilateral lesions won more than other participants did, coupled with more strategic betting and reduced biases, which were attenuated in unilateral patients too. vmPFC may bring contextual information to influence action, which may be suboptimal in some situations.

We find that experts and lay people fared poorly at predicting social and psychological consequences of the pandemic and misperceive what effects it may have already had

Hutcherson, Cendri, Constantine Sharpinskyi, Michael E. W. Varnum, PhD, Amanda M. Rotella, Alexandra Wormley, Louis Tay, and Igor Grossmann. 2021. “The Pandemic Fallacy: Inaccuracy of Social Scientists’ and Lay Judgments About Covid-19’s Societal Consequences in America.” PsyArXiv. February 11. doi:10.31234/osf.io/g8f9s

Abstract: Effective management of global crises relies on expert judgment of their societal effects. How accurate are such judgments? In the spring of 2020, we asked social scientists (N = 717) and lay Americans (N = 394) to make predictions about COVID-19 pandemic-related societal change across social and psychological domains. Six months later we obtained retrospective assessments for the same domains (Nscientists = 270; NlayP = 411) and compared these judgments to objective data to assess estimation accuracy. Social scientists were no more accurate than lay people, neither in prospective nor retrospective judgments. Across studies and samples, estimates of the magnitude of change were off by more than 20% and less than half of participants accurately predicted the direction of changes. Taken together, we find that experts and lay people fared poorly at predicting social and psychological consequences of the pandemic and misperceive what effects it may have already had.

High-income women may prenatally masculinize their sons at the expense of the fitness of their daughters; low-income women may prenatally feminize their daughters at the fitness expense of their sons

Parental income inequality and children’s digit ratio (2D:4D): a ‘Trivers-Willard’ effect on prenatal androgenization? J.T. Manning et al. Journal of Biosocial Science, February 2021. https://www.cambridge.org/core/journals/journal-of-biosocial-science/article/parental-income-inequality-and-childrens-digit-ratio-2d4d-a-triverswillard-effect-on-prenatal-androgenization/CE6A605963BB8D8D921ECBDD4988B01E

Abstract: Income inequality is associated positively with disease prevalence and mortality. Digit ratio (2D:4D) – a negative proxy for prenatal testosterone and a positive correlate of prenatal oestrogen – is related to several diseases. This study examined the association of income inequality (operationalized as relative parental income) and children’s 2D:4D. Participants self-measured finger lengths (2D=index finger, and 4D=ring finger) in a large online survey conducted in July 2005 (the BBC Internet Study) and reported their parents’ income. Children of parents of above-average income had low 2D:4D (high prenatal testosterone, low prenatal oestrogen) while the children of parents of below-average income had high 2D:4D (low prenatal testosterone, high prenatal oestrogen). The effects were significant in the total sample, present among Whites (the largest group in the sample), in the two largest national samples (UK and USA) and were greater for males than females. The findings suggest a Trivers-Willard effect, such that high-income women may prenatally masculinize their sons at the expense of the fitness of their daughters. Women with low income may prenatally feminize their daughters at the fitness expense of their sons. The effect could, in part, explain associations between low income, high 2D:4D (low prenatal testosterone) and some major causes of mortality such as cardiovascular disease.

Keywords 2D:4D Income disparity Trivers-Willard hypothesis

Discussion

The present study found that parental income affects children’s 2D:4D such that below-average income is related to high 2D:4D (feminization of the fetus) and above-average parental income is associated with low 2D:4D (masculinization of the fetus). These findings applied to the total study sample, the most numerous ethnic group in the study (i.e. Whites) and the most numerous national samples (UK and USA). Regarding the total sample, for male children, the effect was present for right and left 2D:4D and for all pairwise comparisons between parental income groups above and below the population average. For female children, the effect was also present on right and left hand 2D:4D and was found in pairwise comparisons between parental income groups above and below the population average. However, the female effects were weaker than the male effects, with only two significant pairwise comparisons for right hand 2D:4D and three for left hand 2D:4D.

The present findings are consistent with the Trivers-Willard hypothesis concerning maternal resources and its links to the influence of the mother’s sex steroids on fetal 2D:4D. Thus, mothers with high income will secrete elevated levels of T relative to E during the 1^st trimester of their pregnancy, i.e. they will masculinize their male and female children. In contrast, women with low income will secrete low levels of T:E in the early stages of pregnancy. This hormonal milieu will feminize their male and female children. That is, high-income mothers will increase the fitness of their sons at the expense of their daughters while low-income mothers will increase the fitness of their daughters at the expense of their sons. Thus, there will be sexually antagonistic effects on the children of both high- and low-income mothers (Manning et al., 2000). There is evidence for an effect of female condition on the production of sex steroids and sexually antagonistic effects of maternal sex steroids on the developing fetus.

There is only weak support for a link between the 2D:4D of women selected at random from the population and their production of T and E (Muller et al., 2011). However, in contrast to women with low income, high-income women may benefit from high levels of nutrition. There may be associations between high income, good nutrition and the production of androgens in women. Elite women athletes with high standards of nutrition and in good condition (with high lean mass and low body fat levels) show negative relationships between their 2D:4D and the breakdown products of T (Eklund et al., 2020) or salivary levels of T (Crewther & Cook, 2019). It appears that women in good condition can, and probably do, secrete elevated levels of T, particularly if they have low 2D:4D.

The Trivers-Willard hypothesis was originally formulated in the context of maternal manipulation of the sex ratio of progeny. There is indeed evidence that masculinized women (with high WHR and/or low 2D:4D) have more sons than feminized women (low WHR and/or high 2D:4D). This may be the result of the deleterious effects of high prenatal T on female fetuses and/or the advantageous effect of high prenatal T on male fetuses (Manning et al., 1996; Singh & Zambarano, 1997; Manning & Bundred, 2001; Kim et al., 2015).

Maternal manipulation of the prenatal sex steroid environment of their children is likely to have later-life health consequences. For example, male children of low-income mothers will be feminized and more prone to several diseases. Prominent among these is likely to be the poverty influenced male-biased burden of cardiovascular diseases. A low income level has been consistently associated with cardiovascular disease, especially in high-income countries. In addition, disparities based on sex (males>females) have been shown in several studies. High 2D:4D in men has been linked to poor outcomes for cardiovascular disease such as early myocardial infarction, high blood pressure, atherosclerotic plaque development, high fibrinogen levels and markers of obesity (Manning & Bundred, 2001; Fink et al., 2006; Lu et al., 2008, 2015; Ozdogmus et al., 2010; Kyriakidis et al., 2010 ; Wu et al., 2013; Manning et al., 2019; Bagepally et al., 2020). Associated with all of these factors is a high level of parental poverty (Kucharska-Newton et al., 2011; Mosquera et al., 2016).

One possible limitation of the present study is that estimates of parental income in the early years of the family are dependent on their children’s recall. Inaccuracies that may result from faulty recall are likely to reduce the Trivers-Willard influence on 2D:4D. Thus, the reported effects may be conservative estimates of maternal influence on offspring 2D:4D. In order to minimize the recall effects of children’s estimates of parental income it is suggested that future studies should also include parental reports of family income.

In conclusion, inequality in parental income may be associated with the 2D:4D of their offspring. Children of parents of above-average income had low 2D:4D (high prenatal T:E) while the children of parents of below-average income had high 2D:4D (low prenatal T:E). The differences in offspring 2D:4D across income groups may arise because of maternal manipulation of sex steroids. Interpreting the findings of the present study through the lens of the Trivers-Willard hypothesis suggests that high-income mothers may masculinize their sons via increased levels of prenatal T. Male reproductive success shows higher variance than female reproductive success. Therefore, the fitness rewards from the sons of high-income mothers are likely to outweigh the deleterious effects of T on the daughters of high-income mothers. In contrast, mothers with low income are expected to feminize their children via increases in prenatal E. The fitness gain from feminized daughters is likely to outweigh the fitness loss of feminized sons. Moreover, the health costs of maternal manipulation of prenatal sex steroids may include hypertension, cardiovascular disease, high levels of fibrinogen and early myocardial infarction and could be focused on the feminized (low T, high E) sons of low-income mothers.