{"paper_id":"285471bd-4f0f-495c-91fd-cfa323535a5d","body_text":"Mechanistic quantification of the effects of genes, environment, and development on human behaviour | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return\"[object Function]\"==o.call(a)}function e(a){return\"string\"==typeof a}function f(){}function g(a){return!a||\"loaded\"==a||\"complete\"==a||\"uninitialized\"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){(\"c\"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){\"img\"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),\"object\"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height=\"0\",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),\"img\"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||\"j\",e(a)?i(\"c\"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName(\"script\")[0],o={}.toString,p=[],q=0,r=\"MozAppearance\"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&\"[object Opera]\"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?\"object\":l?\"script\":\"img\",v=l?\"script\":u,w=Array.isArray||function(a){return\"[object Array]\"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split(\"!\"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split(\"=\"),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(\".\").pop().split(\"?\").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split(\"/\").pop().split(\"?\")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&\"css\"==i.url.split(\".\").pop().split(\"?\").shift()?\"c\":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results Mechanistic quantification of the effects of genes, environment, and development on human behaviour View ORCID Profile Mauricio González-Forero , View ORCID Profile Aida Gómez-Robles doi: https://doi.org/10.1101/2025.03.17.643651 Mauricio González-Forero 1 School of Biology, University of St Andrews , Dyers Brae, St Andrews, KY16 9TH, Fife, UK 2 Konrad Lorenz Institute for Evolution and Cognition Research , Klosterneuburg A-3400, Austria Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Mauricio González-Forero For correspondence: Mauricio.Gonzalez-Forero{at}kli.ac.at Aida Gómez-Robles 3 Department of Anthropology, University College London , 14 Taviton St, London, WC1H 0BW, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Aida Gómez-Robles For correspondence: a.gomez-robles{at}ucl.ac.uk Abstract Full Text Info/History Metrics Supplementary material Data/Code Preview PDF Abstract A long-standing question is to what extent behaviour is determined by genes, environment, or else. Common practice estimates genetic components using heritability, which is correlational, population-level, and variously problematic to measure genetic contribution. Ideally, contributions would be quantified via experimental intervention, which is often infeasible for humans and other organisms, but possible in silico . We formulate mathematical methods to quantify contributions to phenotypic differences in silico , and apply them to an in-silico replica of human skill development and evolution. We find that, in this replica, adult skill level is more reactive to differences in developmental history than to genetic, environmental, or social differences from Homo habilis to H. sapiens . This occurs although, in this replica, adult skill level has evolved an increased reactivity to environmental and social differences, and a decreased reactivity to genetic differences, which is consistent with observation. Our results predict a window of sensitivity at menarche. They also predict that adverse genetic and environmental conditions that would developmentally yield low adult skill level can be compensated by time-specific environmental changes or slight changes in developmental history, potentially yielding outstanding adult skill level. Our analyses suggest a major capability of developmental history to influence adult human skill level. Introduction Adult human behaviour is the product of a complex interplay of genetic, developmental, environmental, and social factors ( Anreiter, Sokolowski, & Sokolowski, 2018 ; Boyce, Sokolowski, & Robinson, 2020 ; Johnson, 2007 ). While the brain expresses thousands of genes ( Zeng et al., 2012 ) and extensive research efforts seek to identify genes contributing to behaviour ( Canela-Xandri, Rawlik, & Tenesa, 2018 ; Sakaue et al., 2021 ), multiple lines of evidence point to the existence of human specialisations for neuroplasticity ( Drennan, Sherwood, & Gómez-Robles, 2026 ; K. R. Rosenberg, 2021 ). Moreover, behaviour is affected by developmental history as evidenced by studies of essentially genetically identical individuals growing in the same environment who develop different behaviours, indicating that developmental noise may be an important source of behavioural variation ( Freund et al., 2013 ; Mitchell, 2018 ; Turkheimer & Waldron, 2000 ). Indeed, up to 50% of variation in human cognitive abilities has been estimated to stem from developmental noise ( Plomin & Kawakami, 2025 ). The high level of plasticity of the human brain helps explain human learning abilities ( van Schaik & Burkart, 2011 ) and makes it possible to develop intervention strategies that can change behaviours over life ( Sparling, Dragomir, Ramey, & Florescu, 2005 ). Observed differences between chimpanzees and humans indicate that increased neuroplasticity evolved in humans after both lineages diverged, together with an overall increase in brain size and a general reorganization of the brain ( Drennan et al., 2026 ). However, disentangling the causes of variation in human behaviour remains challenging. Multiple approaches have been devised to assess causality, but difficulties persist given the commonly observational nature of human data ( Kievit et al., 2018 ; Ota et al., 2025 ; M. D. Rosenberg, Casey, & Holmes, 2018 ; Turkheimer & Waldron, 2000 ). For instance, a common but problematic way to quantify the causes of variation is via heritability: a trait with higher heritability is sometimes interpreted as having a stronger genetic component and being less plastic. Heritability, however, is correlational, depends on population-level phenomena such as genetic, phenotypic, and environmental variation, and does not quantify the causes of a phenotype for a given individual ( Feldman & Lewontin, 1975 ). Moreover, unlike brain size and reorganization, which can be inferred to some extent from endocranial size and shape ( Coqueugniot, Hublin, Veillon, Houët, & Jacob, 2004 ; Gunz et al., 2020 ; Gunz, Neubauer, Maureille, & Hublin, 2010 ), quantifying the relative influence of genetic and environmental factors on behaviour from fossils remains out of reach. Ideally, quantifying the causes of phenotypic variation at the individual-level would be done by determining the effects of experimental intervention, which is a gold-standard to establish causality ( Pearl, 2009 ). For many human behaviours, such interventions would involve infeasible actions such as altering the genetic and environmental make-up of an individual’s past. Although such interventions are often impossible with real individuals, they are possible in silico . Many mathematical or computational models exist that mechanistically describe the development of phenotypes (e.g., mammalian teeth, Salazar-Ciudad & Jernvall, 2010 ; vertebrate digits, Sheth et al., 2012 ; and human skill level, Koster et al., 2020 ). By analysing experimental in-silico interventions in these models, it is possible to quantify the causes of variation in the developed in-silico phenotypes, yielding an inference of the causes of variation in the observed phenotypes in real organisms. In this paper, we use, develop, and apply four methods to quantify the causes of individual-level phenotypic variation when a model of development is available. First, we use the developmental sensitivity of a phenotype ( González-Forero, 2024b ), which quantifies the total effect of marginally changing a variable at a particular point earlier in development on another variable later in development. Second, we use the developmental elasticity of a phenotype ( Caswell, 2019 ), which quantifies the proportional total effect of a variable on another variable. Third, we define developmental reactivity , which quantifies the proportional total effect of a variable on another variable relative to all variables affecting development. Fourth, we define first-order contributions , which quantify the mechanistic (not correlational) genetic and envi-ronmental contributions to the differences between the phenotypes of individuals developed under different conditions. We apply these methods to quantify how much genetic, environmental, developmental, and social factors causally affect the development of different adult human skill levels in a model of human brain size development and evolution ( González-Forero, 2024a ; GonzálezForero, Faulwasser, & Lehmann, 2017; González-Forero & Gardner, 2018 ). We refer to this model as “brain model”, given its original emphasis on modelling brain size evolution. In this model, the brain enables a cognitive ability termed “skill”, defined as an ability of performing some behaviour(s), where some of the brain’s energy consumption is for increasing and maintaining this ability, that is, for “learning” and “memory” (Methods). Models typically make unrealistic simplifications to explore the effect of the modelled factors on the process of interest. This is the case in this model, but we use it because it is able to replicate major patterns of human development and evolution, thereby being able to mechanistically and quantitatively describe a wide array of observed patterns that are often viewed as characteristic of humans. This ability of the model suggests that it may already capture some of the important causes of human phenotypic variation. Yet, other models of human skill development exist (e.g., Koster et al., 2020 ). Although these models do not necessarily include every key aspect of human phenotypic variation, the methods we use can be applied in future studies to modifications of this model or other models to compare their output. The brain model has been previously found to quantitatively and mechanistically recover major aspects of human development and evolution under conditions identified by the model ( Fig. 1 ; González-Forero, 2024a ; González-Forero et al., 2017 ; González-Forero & Gardner, 2018 ; González-Forero & Gómez-Robles, 2025 ). Specifically, it recovers the expansion of adult brain and body sizes from Homo habilis (empirical data from McHenry & Coffing, 2000 ) to H. sapiens scale (empirical data from autopsy hospital reports from Washington, DC, and Bethesda, MD collected between 1964 and 1973; Dekaban & Sadowsky, 1978 ). Simultaneously, the model recovers the evolution of childhood, adolescence, and adulthood periods with timing matching that observed in H. sapiens (empirical data reviewed in Gluckman & Hanson, 2006 ). Moreover, this model identifies a transition to seemingly cumulative culture in the Middle Pleistocene ( González-Forero & Gardner, 2018 ; González-Forero & Gómez-Robles, 2025 ), which is consistent with the archaeological record ( Barham et al., 2023 ; Paige & Perreault, 2024 ; Roebroeks & Villa, 2011 ). Download figure Open in new tab Figure 1: In-silico replica of the evolution and development of hominin brain size, body size, and skill level as obtained by the brain model. ( A ) Green squares show the empirically observed mean adult brain and body sizes in log-log scale for 13 hominin species (data for females when possible). Coloured dots give the recovered evolution of adult brain and body sizes by the brain model, shifting from those of Homo habilis to those of H. sapiens . ( B-E ) Developmental dynamics underlying the trajectory in A : for ( B ) brain size, ( C ) pre-ovulatory ovarian follicle count (in mass units), ( D ) body size, and ( E ) skill level. Panels B,D also show empirically observed cross-sectional mean values of H. sapiens females (black squares), Pan troglodytes females (gray triangles), and adult H. habilis of mixed sexes for brain size and of females for body size (pink stars). One evolutionary time step is the time from mutation to fixation, which would be 11.5 kyr assuming fixation takes 500 generations and generation time is 23 years. All panels are taken with modification from González-Forero (2024a) , which lists references to the data sources. The model also mechanistically recovers a linear relationship between skill level (or cognitive ability) and brain size, which is consistent with observation. Specifically, the model finds that, if skill level and brain size plateau over life, adult skill level and adult brain size are proportional (eqs. 13 and 14 of González-Forero et al., 2017 , briefly rederived in section S.1.3.4 of the supplement here): that is, adult skill level = proportionality factor × adult brain size, where the proportionality factor increases with the metabolic costs of brain maintenance, decreases with the metabolic costs of memory, and increases with the fraction of the brain metabolic rate allocated to the skill considered. This proportionality is not an assumption but a result of the model’s energy conservation analysis, involving fewer assumptions than those made by the whole model and so being relatively general. The proportionality is consistent with brain size and cognitive abilities being positively correlated in primates ( Deaner, Isler, Burkart, & van Schaik, 2007 ; Heldstab et al., 2016 ) and other vertebrates ( MacLean et al., 2014 ). It also agrees with the consistent yet modest positive association between brain size and cognitive ability in humans ( Cox, Ritchie, Fawns-Ritchie, Tucker-Drob, & Deary, 2019 ; Nave, Jung, Linnér, Kable, & Koellinger, 2019 ). This proportionality predicts that some of the variation in the empirically observed relationship between brain size and cognitive ability could be due to variation in the metabolic costs of memory and of brain maintenance and in the fraction of brain metabolic rate allocated to the cognitive ability evaluated. The brain model is relatively complex, which raises concerns that are standard with complex models, specifically, unfalsifiability, underdeterminacy, and overfitting. First, an important concern with models with many parameters is that they may be able to fit any data, thereby being unfalsifiable ( Burnham & Anderson, 2002 ; Dyson, 2004 ) (for an interactive illustration, see https://demonstrations.wolfram.com/FittingAnElephant/ ). This concern is of particular relevance when a model is based on little theory and key parameters are not empirically measured but estimated by model fitting, so such a model has largely no (active) constraints that prevent it from fitting any data. The brain model has many parameters (26 that affect the output) but has a particular form derived from mechanistic considerations, with about half of the parameters having values from empirical estimates (12, including 7 that pertain to metabolic costs, which are thought to be key given the high metabolic cost of brain tissue). This mechanistic form and empirically estimated parameter values appear to make the model actively constrained, a technical term that entails that the model is unable to generate any observation. For instance, the model has so far been unable to recover brain growth in the first year of age, and ontogenetic patterns do not exactly match observation ( Fig. 1 ; González-Forero, 2024a ; González-Forero et al., 2017 ; González-Forero & Gardner, 2018 ; González-Forero & Gómez-Robles, 2025 ). This inability to generate any pattern makes the model rather easily falsified, which is an advantage to identify gaps in understanding ( McElreath, 2018 ). Second, a concern with fitting complex models is that if the model has more parameters than data points, the problem may be underdetermined, which means that multiple parameter values could fit the data equally well. Similarly, this problem is particularly relevant when a model is relatively unconstrained. This problem is relevant here as the brain model has many parameters but is fitted to sparse, fossil data. However, underdeterminacy has not been a problem in the model, as the re-covered adult hominin-brain and body sizes with associated human life history occur only under unique parameter combinations, with fit quickly falling away from them, at least for the region of the parameter space evaluated ( González-Forero & Gardner, 2018 ). Again, this is presumably because the model’s mechanistic basis and empirically estimated parameter values constrain it sufficiently to prevent underdeterminacy. Third, complex models may be overfit, where despite fitting the training data well, they predict out of sample data poorly. This is an important yet still unaddressed challenge for the brain model. Methods exist to estimate overfitting, such as cross-validation, where the model is trained excluding some of the data to evaluate its ability to predict the excluded data ( McElreath, 2020 ). The complexity of the brain model has prevented such analyses but substantial improvements in its runtime make those analyses feasible in the near future ( González-Forero, 2024a ). Overall, at present, this brain model mechanistically, falsifiably, and in a determined manner recovers many, but not all, major patterns of human development and evolution while mechanistically describing skill ontogeny, but the model’s level of overfitting remains to be assessed. Here we use this model to causally quantify in in-silico individuals: 1) the responses of adult skill level to differences experienced earlier in life, 2) how these responses have evolved from H. habilis to H. sapiens , and 3) the genetic and environmental contributions to differences in adult skill level. This analysis finds relatively small genetic and large developmental effects on adult skill level, such that slightly intervening developmental history can compensate for genetic and environmental adversity. The analysis also finds a notable and unexpected window of sensitivity of adult skill to perturbations at the age of menarche and shortly thereafter. Methods Brain model overview The brain model has the following structure. It can be seen as having three major components: development, selection, and evolution. Development The development component ( Fig. 2A ) may be subdivided into two parts: a metabolic part and a behavioural part. Download figure Open in new tab Figure 2: Causal diagram of development in the brain model and measures of the effects of perturbation. ( A ) Causal diagram of development in the model, where an arrow indicates a direct causal effect of a variable on another one or on itself at the next age. Variables are classified into genotypic, phenotypic, environmental/physiological, and social environmental. Genotypic variables are controlled by the individual’s genotype. Phenotypic variables are constructed over development; the developmental history for a given age is defined here as the phenotypic variables at earlier ages. Environmental or physiological variables affect the development of phenotypic variables. Social environmental variables are the skill level (and so are phenotypic variables) of social partners, which affect the development of the focal individual. ( B ) We quantify the effect of perturbation, that is, of altering a variable at a given age, on adult skill level, using four measures: sensitivity d x /d y , which measures total effects, that is, the effect of altering a variable over all causal pathways linking it to the response variable; elasticity ( y / x )d x /d y , which measures the proportional total effects; reactivity, which normalizes elasticity over all the possible interventions over development; and first-order contributions, which quantify the mechanistic contribution to phenotypic differences of genes and environment. Metabolic part The metabolic part considers the body of an individual to be made of three tissues: brain, reproductive, and the remainder referred to as somatic tissue. This body partitioning was considered because of the model’s original interest in brain size evolution and because reproductive tissue enabled modelling fertility, but the model can consider other partitions (e.g., brain regions or digestive system). The metabolic part then describes the growth rates of tissue sizes based on energy conservation; specifically, following West, Brown, and Enquist (2001) , the model assumes that the energy consumed by a tissue equals the energy spent by the tissue on its growth and maintenance, yielding an equation describing the growth rate of the tissue in terms of the fraction of the energy spent by the body on growth that is allocated to the tissue under consideration. The metabolic part also describes the individual’s rate of learning of skills of some kind based on energy conservation; in the metabolic part, it is not specified what the skills are or what they do, so in that sense skills can be interpreted quite generally. The metabolic part obtains an equation describing this learning rate by assuming that some of the energy that is spent by the brain is due to the learning and memory of these skills. This learning rate equation yields the proportionality between adult brain size and adult skill level. The metabolic part enables the model to be parameterised with empirical data of brain and skill metabolic costs. With this empirical parameterisation, the model makes quantitative predictions of brain size and skill level in the units given by the empirically estimated metabolic costs. For instance, empirical estimates of metabolic costs of brain maintenance in units of energy/(mass × time) such as joules/(kilograms × years) lead to predictions of brain size in kilograms at a given age in years. Similarly, empirical estimates of metabolic costs of skill maintenance (memory) in units such as joules/(skill unit × years) lead to predictions of skill level in the skill unit at a given age in years. Although empirical estimates of the metabolic costs of memory of human skills appear to be unavailable, there exist empirical estimates of the metabolic cost of information storage (in bytes) in synapses in Purkinje cells in rats ( Brunel, Hakim, Isope, Nadal, & Barbour, 2004 ; Howarth, Peppiatt-Wildman, & Attwell, 2010 ; Steuber et al., 2007 ). The model has used these estimates as relevant benchmarks for the metabolic cost of memory, which entails that the model predicts skills in information units (bytes). Future empirical estimates of the metabolic costs of memory and learning of specific skills would be helpful to improve the model accuracy. Behavioural part The behavioural part of the model specifies the skills considered and what they do. The model has so far considered “energy-extraction” skills, defined as the capacity of an individual to overcome ecological or social challenges to extract energy from the environment. Ecological challenges refer to the individual having to obtain energy by using their own skills to solve problems posed by the environment excluding social partners. Social challenges can be either cooperative or competitive, where for simplicity and as a first approximation, the model considers these social interactions to be pairwise and only between individuals of the same age. In cooperative social challenges, an individual must use their skills jointly with the skills of another individual of the same age to obtain energy from the environment. In competitive social challenges, an individual must use their skills either individually or in cooperation with another individual of the same to compete with the skills, respectively, of another individual of the same age or with another pair of individuals of the same age. The energy extracted can then be used to grow and maintain the various bodily tissues. The model implements maternal care phenomenologically rather than mechanisti-cally, where the individual’s skills only become necessary for energy extraction as the individual’s age advances. The model so far considers the effect of culture or social learning implicitly, via the shape of the function relating skills to the energy extraction efficiency: the function may have strongly or weakly diminishing returns of increasing skill. Strongly diminishing returns could in principle arise under “non-cumulative” cultural settings where individuals cannot increase their energy extraction efficiency sub-stantially after achieving a high skill level; in contrast, weakly diminishing returns could in principle arise under cumulative cultural settings where individuals can increase their energy extraction efficiency despite already being highly skilled, if they can continue to learn from accumulated knowledge in the population. The model has thus not yet considered a vast array of complexities that are likely to be key to skill development, such as explicit social learning dynamics or social interactions between individuals of different ages. However, the model offers a stepping stone towards further realism. Selection The selection component of the model specifies what selection acts on. To do this, the model makes additionally strong assumptions that were attempted for simplicity as a first approximation. Despite the lack of realism of these assumptions, the model was still able to recover major patterns of human brain and body size development and evolution, so we continue to make them here, again as a first approximation to be relaxed in the future. The selection component assumes that the population is composed of females only that reproduce clonally (i.e., asexually), a standard assumption in life history theory. Further, the model assumes that the fertility of females is proportional to the size of reproductive tissue, defined as the number of mature ovarian follicles, which simplified the mathematical treatment but is consistent with clinical practice ( Broekmans, Faddy, Scheffer, & te Velde, 2004 ; Chang, Chiang, Hsieh, Soong, & Hsu, 1998 ). Additionally, for simplicity and as a first approximation, the model assumes that the probability of surviving from one age to the next is constant; that is, having higher skill levels does not make individuals more likely to survive to the next age. As a result, the selection component entails that there is only selection for reproductive tissue but no selection for cognition, measured as skill level in the model. These assumptions go clearly against standard thinking, but surprisingly a wide array of major patterns of human development and evolution is still recovered. This raises the possibility that conditions that are traditionally regarded as necessary for human evolution (e.g., selection for cognition) might not be as necessary. Evolution The final component of the model is evolution, which allows the model to obtain evolutionary trajectories of brain and body sizes over generations. This component makes additional simplifying assumptions that are not realistic, but again serve as a stepping stone for further re-alism. The model assumes that the fraction of energy spent in growth that is allocated to the growth of each tissue at each age is controlled by a separate genetic locus. Since the model considers three tissues and individuals from birth to 47 years of age, with ten age bins per year, this means that the model considers 1470 loci (each with a continuum of alleles, so each locus can take any real value). Because reproduction is clonal, it is unnecessary to specify how these loci are linked or whether they are in autosomes or sex chromosomes. This assumption of separate loci for each age is a genetic specification of the standard life history approach with genetic traits that vary with age, where the genetic basis of these traits is often not made explicit ( León, 1976 ); this assumption is also mathematically equivalent to assuming that a single locus controls how a genotypic trait varies with age over the life course ( Avila & Lehmann, 2023 ). Given clonal reproduction, the genotypic traits, which control the energy allocation to tissue growth, are passed on to offspring without alteration barring rare mutation. These are genetic idealisations that are not realistic, but are standard and as before we use them here for simplicity as a first approximation. The model assumes that evolution occurs by mutationfixation epochs: at the start of each epoch, all individuals have the same genotypic traits and the population attains carrying capacity, then a mutation of small effect on the genotypic traits arises, and finally, mutants either vanish or spread to fixation and the population attains a new carrying capacity. These stylised assumptions of mutation-fixation epochs are again made for simplicity, are common in population genetics modelling, and are standard in adaptive dynamics ( Dieckmann & Law, 1996 ; Metz, Geritz, Meszéna, Jacobs, & van Heerwaarden, 1996 ). These assumptions of evolution by mutation-fixation epochs are often found to yield approximately correct evolutionary trajectories if mutation is rare and of small effect ( Champagnat, 2006 ). As the genetic loci evolve, the phenotypic traits, namely the sizes of the various bodily tissues and skill level across the life course evolve in response. Under the parameter values that recover human brain size evolution, the loci evolve such that mature ovarian follicles only start being produced after a long period of childhood ( Fig. 1C , red). We refer to the age at which mature ovarian follicles are first produced as the age at menarche, which is an emergent feature of the model rather than being a predefined parameter. The evolution component of the model allows it to show that the recovered hominin brain expansion occurs because a challenging ecology and seemingly cumulative culture affect development such that brain size and the only tissue under selection in the model, namely, reproductive tissue, become genetically correlated (GonzálezForero, 2024a). That is, the human brain size evolves in the model as a spandrel, or by-product of selection for reproductive tissues ( González-Forero & Gómez-Robles, 2025 ). This conclusion is surprising but not definitive as further comparisons with model variations are necessary, particularly where selection for cognition is al-lowed (e.g., where survival is a function of skill level). However, this surprising conclusion is unexpectedly consistent with empirical evidence of particularly high and causally relevant genetic correlations between brain related traits and reproductive traits in human females ( Molz et al., 2024 ), human males ( Bush & Goriely, 2024 , 2025 ; Matos, Publicover, Castro, Esteves, & Fardilha, 2021 ), and fish ( Buechel, Booksmythe, Kotrschal, Jennions, & Kolm, 2016 ; Corral-López et al., 2017 ; Kotrschal et al., 2015 ). Specifically, it has been found that the shape of the braincase in human females is genetically correlated with female reproductive tissues ( Molz et al., 2024 ), that there is a large overlap in genes associated with cell proliferation in the human brain and testes ( Bush & Goriely, 2024 , 2025 ; Matos et al., 2021 ), and that artificially selecting for large brains in fish causes an associated evolution of enlarged male genitalia and other sexual traits ( Buechel et al., 2016 ; Corral-López et al., 2017 ; Kotrschal et al., 2015 ). Developmental sensitivity We seek to quantify the extent to which genes, environment, and any other factors involved in development causally affect the developed skill level across the recovered hominin brain expansion. Causal effects can be estimated using sensitivity analyses of a model ( Caswell, 2019 ). Such causal effects can be computed with derivatives, provided that such derivatives are causal ( Henshaw, Morrissey, & Jones, 2020 ). For instance, if a variable x causally affects y = f ( x ) for some function f but y does not causally affect x , then causal derivatives of x with respect to y must be zero even if f is invertible ( Lehtonen & Otsuka, 2023 ). We compute developmental sensitivities using recently derived formulas for the sensitivity of a recurrence ( González-Forero, 2024b ), which give causal derivatives by considering a mechanistic description of the system by means of recurrence equations, allowing for arbitrarily complex and non-linear causal feedbacks. This causality from derivatives is achieved by computing the effects of a perturbation on the recurrence from the time of the perturbation, specifically, by applying a perturbation at some time and following the effects of the perturbation from that time onwards ( González-Forero, 2024b ). This approach considers the causal arrow of time, such that perturbations can only causally affect the future, but not the past. These formulas also include causal feedbacks (they all depend on a matrix describing developmental feedbacks; Eq. S5): that is, when a perturbation occurs at a given age, it may affect phenotypes that affect them-selves as age advances (via all direct and indirect loops in Fig. 2A ). For the brain model, we analyse the sensitivity of the skill level of an in-silico individual at a given age to changes earlier in development, and how such sensitivity has changed over evolution ( Fig. 2B ). We compute the sensitivity of skill level to developmentally earlier point perturbations in genotypic traits (i.e., the energy allocation efforts to grow each tissue), phenotypic traits (i.e., brain size, somatic size, mature ovarian follicle count, and skill level), the skill level of an average social partner, or the parameters affecting development; for short, we say that the parameters affecting development describe environment or physiology, as these parameters refer to environmental, metabolic, social, and maternal care properties. For instance, we compute the effect on the skill level of a 40-year-old individual caused by a genotypic change that marginally changed her energy allocation to brain growth at 1 year of age but only at that 1 year of age. Similarly, we also compute the effect on the skill level of a 40-year-old caused by a phenotypic change that marginally changed, for instance, her body size when she was 2 years old. Our approach does not need to specify how those interventions occur in practice, but only predicts their effects if the interventions were to occur (as discussed in p. 6 of Caswell, 2019 ). Although it may appear that the effects of an intervention would be completely diluted by random life events as the time between the intervention and the evaluated age increases, in a later section we introduce realistic levels of randomness finding that the effects of the intervention do not necessarily wane and instead may increase as the effects of the perturbation propagate over development. In quantifying developmental sensitivity, we only consider the effects of interventions on factors directly affecting the development of the focal individual, rather than interventions on factors that indirectly affect the focal individual’s development. For instance, in the brain model, the focal individual’s development directly depends on the skill level of social partners because of cooperation and competition, but does not directly depend on the body size, brain size, or environmental conditions experienced by social partners, which affect the social partner’s skill level and so indirectly affect the focal individual’s development. Thus, our developmental sensitivity calculations only consider the effects of intervening the skill level of social partners, but not the effects of perturbing the brain size, body size, or environmental conditions of social partners. The model at present is likely to underestimate the developmental sensitivity to social interactions. This is because social learning has only been implicitly incorporated in the model. Thus, intervening the skill level of social partners only affects the focal individual’s development because of cooperation and competition for energy extraction, rather than via social learning. However, we do consider some aspects of social learning implicitly, specifically by intervening parameters whose value may be partly due to culture (i.e., skill effectiveness at energy extraction and environmental difficulty, which affect the shape of energy extraction efficiency vs skill level). Developmental elasticity The units of the developmental sensitivities depend on the perturbed and response variables. To compare sensitivities despite their different units, a standard approach is to compute the elasticity ( Caswell, 2019 ), which is the proportional sensitivity, so is dimensionless and is defined only for positive variables. As genotypic traits can take negative values (e.g., defined as deviations from a reference value), we compute a slight modification of elasticity in this situation, normalising by the absolute value of the genotypic traits. Thus, to compare sensitivities to different types of variables, we analyse such developmental elasticities of skill level over the evolutionary brain expansion from H. habilis to H. sapiens as recovered by the brain model. We refer to the total elasticity of a phenotype at a given age as such phenotype’s elasticity to marginally changing all the variables at all ages (denoted by the vector t ) directly affecting development. Developmental reactivity To quantify in a concise way the relevance of different types of variables for a phenotype of interest, we use the following quantity. We define the developmental reactivity R ζ t k j of a phenotype k at age j to intervening a vector ζ as the elasticity of the phenotype at that age to intervening that vector ζ divided by the total elasticity of the phenotype at that age. For instance, when the intervened vector is all the genotypic traits (denoted by y ), then the reactivity to genotypic traits R yt k j gives the contribution of genotype perturbations to the total elasticity (similarly, R xt k j , , and R ϵ t k j denote, respectively, the reactivity to phenotypic traits, social traits, and environment or physiology, where ϵ is the vector of all parameters affecting development). Reactivities may be negative but add up to 1, so reactivities offer a convenient summary of the developmental sensitivities ( R tt k j = 1, where t is the vector made by the genotypic y , phenotypic x , parameter ϵ , and social variables affecting development). Although each reactivity may be interpreted as the contribution of each variable type to the total elasticity of skill level at a given age, such contributions are not independent (similarly to Caswell, 2001 , p. 229). The relative robustness (or canalisation) of a phenotype to perturbing a variable can be measured by one minus reactivity. Reactivity and robustness defined in this manner are individual-level properties rather than population-level quantities such as heritability. However, note that the reactivity to genetic change is not “the genetic component” of a trait, similarly to heritability which is not the genetic component of a trait although this is a common misinterpretation. Instead, the reactivity to genetic change measures to what extent a phenotype responds to genetic change, relative to all other possible changes in development. For other measures of robustness, see Félix and Barkoulas (2015) and Kitano (2007) . First-order contributions to developed phenotypic differences We consider a situation where two individuals have different phenotypes and there is an ethically warranted interest in quantifying the contributions of genes and environment to these individuals’ different phenotypes. These contributions can be quantified in an in-silico model of development, by combining the developmental sensitivities above that use formulas for the sensitivity of a recurrence ( González-Forero, 2024b ) with the logic of Caswell’s life table response experiment method ( Caswell, 1989 ) (and p. 261 of Caswell, 2001 ). In the supplement, we show that by taking a first-order approximation of the phenotype of one individual around the conditions experienced by the other individual, we obtain an approximated expression that partitions the difference in phenotypes between the two individuals into the contribution from the genetic and environmental conditions experienced (Eqs. S13-S15). As this is a first-order approximation, this quantification is more accurate when the compared individuals experience similar genes and environments. We apply this expression to quantify the contributions of genes and environment to the differences in skill level between in-silico individuals having average bottom and top quantile adult skill levels. To generate empirically informed bottom and top skill levels, we introduce levels of genetic and environmental variation using empirical estimates of heritability as follows. Heritability The above notions can be quantified at the individual level and so do not require consideration of population variability, except that first-order contributions require that the individuals compared have different phenotypes. In contrast, common approaches to quantify genetic and environmental contributions rely on heritability, which requires population variability. To compare our approach to heritability-based approaches and to apply our firstorder contributions approach to empirically informed variation in skill levels, we proceed as follows. First, we introduce genotypic and environmental variation around the average in-silico individual recovered in the Homo sapiens scenario (i.e., yellow dot in Fig. 1A and red lines in Fig. 1B-E ). This average individual has genotypic traits and experiences an environmental difficulty α . A higher environmental difficulty means that individuals need a higher skill level to extract energy from the environment under ecological or cooperative challenges. We independently sample ten thousand values for both the genotypic traits and environmental difficulty around these average values ( and α ) from a normal distribution with variance and , respectively. For each sampled genotypic traits and environmental difficulty, we develop an in-silico individual having such genotype and environment. We thus obtain a population of ten thousand insilico individuals having different phenotypes distributed around the average individual that was recovered by the brain model under the Homo sapiens scenario in Fig. 1 . Second, we calculate the heritability of skill level over the life course among the in-silico individuals that develop under this genotypic and environmental variation. We use the standard definition of narrow sense heritability, namely, the ratio of additive genetic variance and phenotypic variance. The additive genetic variance is the variance in breeding value, defined as the best linear prediction (via leasts squares) of the phenotype from the genes. We obtain skill heritabilities that depend on the genotypic and environmental variance in the population ( and ). Third, to obtain empirically relevant first-order contributions, we identify empirically informed levels of genotypic and environmental variation. We do this by nonexhaustively determining values of this variation that yield a heritability of skill level in young adults that matches observed heritabilities of general cognitive ability in young human adults ( Haworth et al., 2010 ). Our exploration of the genotypic and environmental variation is non-exhaustive, so it is possible that other values of this variation also recover observed values of general cognitive ability in young human adults. This comparison with the heritability of general cognitive ability does not assume that the heritability of skill level is the same as the heritability of general cognitive ability in real people. Instead, since we are unaware of empirical estimates of the heritability of skill level, particularly at multiple stages of life, we do this comparison to compare model outputs with the most relevant empirical data we are aware of. This allows us to roughly infer genotypic and environmental variances. If empirical estimates for skill heritability in young adults were available and substantially different (e.g., much lower than for general cognitive ability), the model results would be unaffected but the estimated genetic and environmental variances and first-order contributions would be different. These differences, however, would not qualitatively affect the key result below that developmental history has substantial importance, which is an individual-level result unaffected by these variances. Results Skill is little elastic to genetic change We find that adult skill level is little elastic to change in genotypic traits in this in-silico replica of hominin evolution. With H. habilis brain and body sizes, perturbing genotypic traits at any age typically has small effects on the developed adult skill level ( Fig. 3A ; the computer code that generated all the figures is in González-Forero, 2026 ). Sensitive periods to genetic change occur early in development and coincide with the onset of reproductive ability, or menarche defined as the developmental emergence of mature ovarian follicles ( Fig. 1C , blue). With H. sapiens brain and body sizes, perturbing genotypic traits also has small effects on adult skill level, but perturbing genotypic traits around the age of menarche (approximately 12 yr; Fig. 1C , red) has more pronounced effects than at other ages ( Fig. 3E ). Download figure Open in new tab Figure 3: Adult skill level is most elastic to change in developmental history. Elasticity of adult skill level to ( A,E ) genotypic, ( B,F ) phenotypic, ( C,G ) environmental or physiological, and ( D,H ) social change: plots give the change in skill level at 40 yr when the indicated variable is marginally increased at the age and only at the age in the horizontal axis. Panels A-D are for individuals with brain and body sizes of H. habilis scale (i.e., for blue dots in Fig. 1B-E ). Panels E-H are for individuals with brain and body sizes of H. sapiens scale (i.e., for red dots in Fig. 1B-E ). In B and F , the elasticity after 40 yr is zero because the skill level at 40 yr is not affected by changes implemented in the future. The elasticity to phenotypic change is largest ( B,F ), followed by the elasticity to environmental ( C,G ), genetic ( A ), and social ( D,H ) change (see vertical axes). x b a : brain size at age a, x r a : mature ovarian follicle count at age a, x s a : somatic tissue size at age a, x k a : skill level at age a, y b a : genetically controlled effort for brain growth at age a, y r a : genetically controlled effort for production of mature ovarian follicles at age a, y s a : genetically controlled effort for somatic growth at age a, α : environmental difficulty, γ : skill effectiveness, ϕ 1 : maternal care at birth, ϕ r : rate of maternal care decrease, : brain size of social partners at age a , : mature ovarian follicle count of social partners at age a , : somatic tissue size of social partners at age a , and : skill level of social partners at age a . Skill is more elastic to developmental change Adult skill level is more elastic to change in developmentally earlier phenotypes, exceeding the elasticity to genetic change by two orders of magnitude with either H. habilis or H. sapiens brain and body sizes ( Fig. 3B,F ). This is unexpected as both genotypic and phenotypic change induce developmental feedback (Eqs. S5 and S6), and both are implemented at a single age only. The analytical expressions of the sensitivity of a recurrence indicate that the sensitivities to genotypic change are relatively small because the direct effects of phenotypes on phenotypes at the next age are mostly positive (Eq. S45; originally derived in González-Forero, 2024a ), whereas the direct effects of genotype on phenotypes at the next age are often negative due to trade-offs in energy allocation to growth of the various tissues (Eq. S50; originally derived in González-Forero, 2024a ). These positive and negative direct effects mean that the total effect of genotype on phenotype may be positive via a path on the causal diagram but negative via another one, and the two effects cancel each other yielding little total genotypic effect on phenotype. With H. habilis brain and body sizes, adult skill level is most elastic to perturbation in skill level or brain size at earlier ages, with a window of elasticity to change in body size ( Fig. 3B ) confined to the short period of brain growth ( Fig. 1B ). With H. sapiens brain and body sizes, adult skill level is most elastic to perturbation in: body size in infancy and early childhood, skill level in mid-childhood, and brain size during most of late childhood and adult-hood ( Fig. 3F ). During adolescence, interventions on variables other than brain size have essentially no proportional effect on adult skill level in the model. Here adolescence is defined as the period where the individual is reproductively capable but keeps growing (approximately 12-19 yrs; Fig. 1C,D ). This inelasticity of adult skill to adolescent perturbation arises seemingly because, by adolescence, brain growth has ceased in the model, as perturbations occurring after brain growth arrest have vanishing effects on adult skill level except if they are on brain or skill. Skill is intermediately elastic to environmental change Altering the environment and physiology has moderate proportional effects on adult skill. With H. habilis brain and body sizes, increasing the environmental difficulty (i.e., individuals need higher skill level to extract energy) has mostly negative proportional effects on adult skill, with a marginally positive proportional effect of environmental difficulty at menarche ( Fig. 3C , blue). Instead, with H. sapiens brain and body sizes, interventions making the environment more difficult during the year after menarche but less difficult in the subsequent year enhance adult skill level ( Fig. 3G , blue). The effectiveness of skills at energy extraction has largely opposite proportional effects to those of environmental difficulty ( Fig. 3C,G , pink). Either increasing maternal care or decreasing the rate of maternal care decrease early in childhood has positive effects on adult skill level ( Fig. 3C,G , orange and green). Social interactions also have moderate developmental effects. With H. habilis brain and body sizes, increasing the social partners’ skill level typically has positive proportional effects on the focal individual’s skill, particularly around menarche ( Fig. 3D ). With H. sapiens brain and body sizes, increasing the skill level of social partners with whom the individual interacts at menarche has marked negative proportional effects on adult skill level ( Fig. 3H , green). Instead, increasing the skill level of social partners with whom the individual interacts at other ages typically has positive proportional effects on adult skill level ( Fig. 3H , green). Evolution of developmental reactivity from Homo habilis to H. sapiens The genotypic reactivity of skill level is small and has decreased over evolution. With brain and body sizes of H. habilis scale, the small reactivity of skill level to genotypic change peaks at around 5 yr ( Fig. 4A , blue). Over evolution to H. sapiens brain and body sizes, genotypic reactivity decreases further, with a peak at around 18 yr ( Fig. 4A , red). Download figure Open in new tab Figure 4: Evolution of developmental reactivity of skill level and brain size over hominin brain expansion. ( A ) Reactivity of skill level to genotypic change; for instance, the red dot at the age a = 10 yr gives the response of a 10-year-old’s skill level to marginally increasing all genotypic traits across all ages relative to changing all variables developmentally affecting skill level at the evolutionary time where adult brain and body sizes are of H. sapiens scale. The blue dot at the same age is for the evolutionary time when adult brain and body sizes are of H. habilis scale. Reactivity of skill level to ( B ) phenotypic, ( C ) environmental or physiological, and ( D ) social change. Skill level reactivity to genotypic and phenotypic change has decreased over evolution ( A, B ) and that to environmental, physiological, and social change has increased over evolution ( C, D ). Adult skill level remains most reactive to phenotypic change ( B ), environmental or physiological change ( C ), genotypic change ( A ), and social change ( D ) in decreasing order (see vertical axes). ( E-H ) Brain size reactivities have similar patterns. ( I ) Pie charts illustrating skill level reactivities in A-D at H. sapiens ( τ = 500) and H. habilis ( τ = 1) brain and body sizes for some ages. The reactivity of skill level to developmental history is much higher despite having decreased over evolution. With H. habilis brain and body sizes, skill level throughout life reacts strongly to phenotypic change, and this reactivity increases with age ( Fig. 4B , blue). Over evolution to H. sapiens brain and body sizes, the reactivity of skill level to developmental history decreases but remains substantial still increasing with age ( Fig. 4B , red). The reactivity of skill level to environmental or physiological change is moderate and has increased over evolution. With H. habilis brain and body sizes, skill level reacts less strongly to environmental or physiological change than to developmental history particularly after childhood, and this reactivity decreases with age ( Fig. 4C , blue). Over the evolution to H. sapiens brain and body sizes, the reactivity of skill level to environmental or physiological change increases, but still decreases with age ( Fig. 4C , red). The reactivity of skill level to social change is very small, but it increases over evolution. With H. habilis brain and body sizes, skill level reacts slightly to changes in the skill level of social partners, with a developmental peak at around 5 yr ( Fig. 4D , blue). Over evolution to H. sapiens , the social reactivity of skill level more than doubles, with a developmental peak at around 9 yr ( Fig. 4D , red). However, for brain and body sizes of H. sapiens scale, the reactivity of skill level to social change is three orders of magnitude smaller than to phenotypic, environmental or physiological change, and one order of magnitude smaller than to genotypic change ( Fig. 4A-D , red). Such small reactivity to social change may result partly because we only consider the effects of perturbing variables that directly affect development (so we do not consider the effects of perturbing variables in social partners that then affect the focal individual) and because we do not model social learning dynamics explicitly although we model social learning implicitly via the shape of the energy extraction efficiency vs skill level. Some of these reactivities are summarised in Fig. 4I . The sensitivities and elasticities underlying the reactivities presented are shown in Figs. S1-S16. The reactivities of the mature ovarian follicle count and somatic tissue size over the recovered human brain expansion are shown in Fig. S17. Our results show a similar overall pattern of evolutionary change in the reactivity of skill level and brain size, due to the proportionality between adult brain size and adult skill level in the model ( Fig. 4 ). Identifying interventions that compensate for adversity We now introduce genetic and environmental variation around the in-silico H. sapiens means. Specifically, we explored a few values for the genotypic and environmental variances ( and ; Fig. S18) and use values that yield a heritability of skill level in young adults approximately matching that empirically observed for general cognitive ability in young adults ( Haworth et al., 2010 ) ( and ; Fig. 5H ). These variances cause some individuals to develop low or high adult skill level ( Fig. 5A-D ), and a predicted skill heritability that increases with age ( Fig. 5H ). This is consistent with the empirically observed increase in heritability of general cognitive ability in humans ( Haworth et al., 2010 ). The predicted increasing heritability of skill level with age contrasts with the decreasing reactivity of skill level to genetic change with age. This exemplifies that the two quantities measure different properties: heritability refers to realised variation, whereas reactivity refers to what would happen due to intervention. The predicted heritabilities of adult brain and body sizes ( Fig. 5E,G ) also fall within empirically estimated ranges: 66-97% for total brain volume (Jiska S. Peper, Boomsma, Kahn, & Pol, 2007) and 47-90% for body mass index ( Elks et al., 2012 ), although the empirically estimated heritability of body mass index appears to be larger in childhood than in adulthood ( Elks et al., 2012 ). Download figure Open in new tab Figure 5: Simulating genetic and environmental adversity. For each age, we draw 10000 genotypic traits and environmental difficulties, normally distributed (variance and , respectively) around their means given by the H. sapiens scenario in Fig. 1 , under which we develop 10000 individuals. ( A-D ) Phenotypes of individuals in the top (dark pink) and bottom (dark blue) 1000th quantile of adult skill level (with respective means in light pink and light blue). Bottom-quantile individuals develop ( A ) smaller adult brain size, ( B ) higher fertility earlier and lower fertility later in life, ( C ) smaller adult body size, and ( D ) smaller adult skill level than the mean (black). ( E-H ) Resulting heitabilities across the 10000 individuals; these heritabilities depend on and (Fig. S18), whose values are chosen so the heritability of skill level of young adults ( H , black at 17 yr) roughly matches the heritability of general cognitive ability of young adults ( H , gray at 17 yr) empirically estimated by Haworth et al. (2010) . First-order contribution of ( I ) genes and ( J ) environment to the skill-level difference between the average individual among the top (light pink) or bottom (light blue) 1000-th quantiles in adult skill level and the population-average individual, showing that genes contribute about twice more to the simulated differences than the environment. ( K ) Environment experienced by the population average individual (black), top (dark pink) and bottom (dark blue) quantile individuals, and by respective mean top (light pink) and bottom (light blue) quantile individuals. Next, we quantify the causal, mechanistic contribution of genes and environment to the simulated differences in adult skill level. For the genotypic and environmental variation used, we find that genes causally contribute over twice as much as the environment to the developed differences in adult skill level between average top- and bottom-skilled in-silico individuals (i.e., the genes of the average top-skilled individual increase their adult skill level by ≈ 0.2 TB, whereas the environment increase it by ≈ 0.09 TB; Fig. 5I,J , pink at 40 yr—the genes and environment of the average bottom-skilled individual decrease their adult skill level by similar magnitudes; Fig. 5I,J , blue). This further exemplifies that heritability, reactivity, and genetic contributions measure different properties: reactivity refers to what would happen after intervention, whereas both heritability and genetic contributions refer to realised variation; however, heritability refers to the fraction of phenotypic variation that is additive genetic, whereas genetic contributions refer to how much phenotypic differences are caused by differences in genes. The first-order contribution of genes to the differences in skill level between average top- and bottom-skilled individuals during childhood is nearly zero as skill differences only become appreciable after menarche ( Fig. 5I ). The first-order contribution of the environment to the differences in skill level becomes appreciable earlier in childhood, partly due to the small contribution in adulthood. Subsequently, we identify interventions that compensate for the simulated adversity. We implement environmental interventions in the direction of predicted steepest increase in adult skill level, given by the sensitivity of adult skill to environmental change ( Fig. 6A , green). These interventions make the environment less difficult during childhood, but more difficult at the ages of sharpest brain growth and of menarche ( Fig. 6A , green). Under the intervened environment, bottom-skilled individuals, who have genes that are detrimental to their skill, develop higher skill level than otherwise top-skilled individuals ( Fig. 6B-E , green). Download figure Open in new tab Figure 6: Slight interventions in developmental history compensate for genetic and environmental adversity. ( A ) Environmental difficulty on average: across the population (black), among the bottom (light blue) and top (light pink) 1000th quantile of adult skill level, and among the bottom 1000th quantile of adult skill level facing environmental intervention (green). The environment is intervened in the direction of steepest increase of adult skill level (i.e., the environment in green is that in blue plus the sensitivity of adult skill level with respect to environmental difficulty). ( B-E ) Corresponding phenotypes developed. Under the intervened environment, bottom-quantile individuals (green) develop ( B ) larger adult brain size, ( C ) average fertility, ( D ) average adult body size, and ( E ) higher adult skill level than top-quantile individuals (pink). ( F ) Slight intervention in skill level in the direction of steepest increase of adult skill level (skill level at each age is increased by 0.01 times the sensitivity of adult skill level with respect to skill level at that age). ( G-J ) Corresponding phenotypes developed. Bottom-quantile individuals (green), who have genes detrimental to adult skill, develop ( G ) larger adult brain size, ( H ) high early fertility, ( I ) slightly larger adult body size, and ( J ) substantially higher adult skill level than top-quantile individuals (pink). Now, rather than intervening the environment, we slightly intervene skill level earlier in life in the direction of predicted steepest increase in adult skill level, given by the sensitivity of adult skill to earlier skill level ( Fig. 6F ). These slight interventions in skill produce larger compensation in adult skill level than the environmental interventions ( Fig. 6G-J ). This stronger compensation in skill requires slighter interventions because of the higher elasticity of skill level to developmental history than to the environment. These interventions are specifically to increase skill level during childhood, at the age of menarche, and as the assessed adult age approaches ( Fig. 6F ). Interventions in skill level in the same direction but of greater magnitude disrupt development (Fig. S19). In practice, these interventions may occur exogenously, for instance, via a supportive learning environment, or endogenously, via developmental noise in brain connectivity. Discussion Complex phenotypes with a high heritability are traditionally interpreted as being largely genetically determined. In contrast, our results illustrate a limited ability of heritability to quantify genetic determinacy and provide alternative measures of it. Our findings indicate that adult human skill level, even with high heritability, may be much more sensitive to the individual’s developmental history than to other factors, including genetic change, environmental change, and changes in the skill level of social partners. Developmental history is here captured by the phenotypes, which store the effects that all developmentally relevant factors (genotype, environment, social context, developmentally earlier phenotypes, and how they all interact during development) have had on the individual until a given age. Our finding that the genotype has limited effects on adult skill level is surprising but consistent with recent findings of very small genetic effects on complex phenotypes in humans, where confounders have been increasingly removed from classic estimates of larger effects ( Berg et al., 2019 ; Sohail et al., 2019 ). Specifically, recent estimates for cognitive performance find a direct genetic effect heritability of 0.233 and a polygenic index with a coefficient of determination of 0.002 (Supplementary Tables 5 and 9 of Tan et al., 2024 ). Our finding that developmental history has the strongest effects on adult skill level is agnostic to what factors change developmental history. Those factors may be endogenous, such as developmental noise ( Freund et al., 2013 ; Mitchell, 2018 ; Plomin & Kawakami, 2025 ; Turkheimer & Waldron, 2000 ), or exogenous, such as targeted interventions ( Sparling et al., 2005 ). The strong sensitivity to developmental history entails that intervening it can be beneficial if done sufficiently slightly but detrimental if done too strongly. This highlights the need of predictive approaches to identify desired interventions. Our results indicate that developmental drivers of cognition are not necessarily reliable indicators of its evolutionary drivers. Indeed, we find that developmental effects can be opposite to evolutionary effects. Specifically, increasing environmental difficulty often decreases skill level over development, whereas increasing environmental difficulty increases skill level over evolution ( Figs. 6 and V of González-Forero et al., 2017 ). This contrasts with previous expectations that developmental drivers of cognition may inform about its evolutionary drivers ( Ashton, Ridley, Edwards, & Thornton, 2018 ). Our results indicate that developmental history and environmental change have an opposite age-specific pattern in influencing skill level. In early life in H. sapiens (up to approximately 10 yr), skill level is influenced by developmental history and environmental variation in roughly equal proportions (approximately 50%) ( Fig. 4B,C ). As individuals age, skill level becomes increasingly reactive to developmental history and insensitive to environmental change, with only 20% of skill-level reactivity to environmental change at 40 yr. Our results suggest that the predominant role of developmental history in affecting skill level may have been shared with many Homo species. According to our results, the proportion of factors influencing skill level in H. sapiens would have been similar 1.1 Myrs after the origin of the genus Homo , roughly corresponding with the H. erectus time period (at evolutionary time τ = 100; Fig. 4 ). We find a reduction in the reactivity to genetic change over evolution and an increase in the reactivity to non-genetic factors, which is consistent with the view that humans have evolved increased plasticity in skill level and brain anatomy ( Drennan et al., 2026 ; Gómez-Robles, Hopkins, Schapiro, & Sherwood, 2015 ; Gómez-Robles, Hopkins, & Sherwood, 2013 ; Hrvoj-Mihic, Bienvenu, Stefanacci, Muotri, & Semendeferi, 2013 ; Snell-Rood, 2013 ). Our results are also consistent with the view that a protracted brain development in H. sapiens allows for stronger environmental influences ( Gómez-Robles, Nicolaou, Smaers, & Sherwood, 2024 ). At H. habilis brain size, skill level substantially reacts to environmental and social factors particularly up to 5 yr, roughly coinciding with the estimated weaning age ( Robson & Wood, 2008 ). Skill level in early Homo species would then be comparatively little influenced by environmental factors after weaning, which is consistent with findings that many primate species reach adult skill levels around weaning age due to the lack of post-weaning provisioning ( Heldstab, Isler, Schuppli, & van Schaik, 2020 ). However, at H. sapiens brain size, environmental and, to a lesser extent, social factors can influence skill level until almost 15 yr, well into adolescence. This finding is consistent with previous studies indicating late brain maturation in humans ( Miller et al., 2012 ; Sotiras et al., 2017 ) associated with late acquisition of adult skill levels ( Kaplan, Hill, Lancaster, & Hurtado, 2000 ; Schuppli, Isler, & van Schaik, 2012 ). That is, H. sapiens brains and skill levels can be strongly influenced by environmental factors at an age at which H. habilis skill level would have been largely insensitive to influences other than developmental history ( Fig. 4I ). Our results also find an unexpected window of sensitivity at the age of menarche and shortly thereafter. This finding is partly consistent with empirical work documenting adolescence as a sensitive period in mammals and birds ( Sachser, Hennessy, & Kaiser, 2018 ; Sachser, Zimmermann, Hennessy, & Kaiser, 2020 ), but our predicted window of sensitivity at menarche quickly closes in early adolescence. Previous work has evaluated whether sensitive periods are adaptive ( Sachser et al., 2018 ) and whether they evolve in response to an organism’s uncertainty about the future ( Walasek, Panchanathan, & Frankenhuis, 2025 ). The sensitive periods we recover do not evolve because of selection for them, but are a byproduct of the developmental process. This is because the model so far only involves selection for mature ovarian follicles, so there is no selection for sensitive periods in the model. This does not mean that the recovered sensitive periods are maladaptive in the model, only that they are not an adaptation under the textbook definition of adaptation as a trait that evolved because of selection for the trait ( Futuyma & Kirkpatrick, 2017 ). Overall, our results suggest a predominant role of de-velopmental history in shaping adult skill level in humans and uncover opportunities for predicting interventions with specific timing and magnitude to mitigate and overcome adverse conditions that would otherwise yield detrimental developmental outcomes. Author Contributions MGF conceived the work and did the mathematical and computational analyses. MGF and AGR analysed the results and wrote the paper. Financial Support MGF was partly funded by an European Research Council Consolidator Grant to A. Gardner (grant no. 771387). AGR is supported by a Royal Society-Leverhulme Trust Senior Research Fellowship (SRF\\R1\\241038). Conflicts of Interest MGF and AGR declare none. Research Transparency and Reproducibility No data were collected in this study. All data used were previously published in references provided in the text. All computer codes used to obtain the results are available at https://doi.org/10.5281/zenodo.19881503 ( González-Forero, 2026 ). Acknowledgements MGF thanks Andy Gardner for continued support that led to this work. We thank Elisabeth Zimmermann and four anonymous reviewers for comments that helped to greatly improve the presentation. Funder Information Declared European Research Council , 771387 Royal Society, https://ror.org/03wnrjx87 , SRF\\R1\\241038 Footnotes This is an extensively revised version to improve the presentation. Results are the same. https://doi.org/10.5281/zenodo.19881503 References ↵ Anreiter , I. , Sokolowski , H. M. , & Sokolowski , M. B. ( 2018 ). Gene–environment interplay and individual differences in behavior . Mind Brain Educ ., 12 , 200 – 211 . OpenUrl ↵ Ashton , B. J. , Ridley , A. R. , Edwards , E. K. , & Thornton , A. ( 2018 ). Cognitive performance is linked to group size and affects fitness in australian magpies . Nature , 554 , 364 – 367 . OpenUrl CrossRef PubMed ↵ Avila , P. , & Lehmann , L. ( 2023 ). Life history and deleterious mutation rate coevolution . J. Theor. Biol ., 573 , 111598 . OpenUrl CrossRef PubMed ↵ Barham , L. , Duller , G. , Candy , I. , Scott , C. , Cartwright , C. , Peterson , J. , … Nkombwe , P. ( 2023 ). Evidence for the earliest structural use of wood at least 476,000 years ago . Nature , 622 , 107 – 111 . OpenUrl CrossRef PubMed ↵ Berg , J. J. , et al. ( 2019 ). Reduced signal for polygenic adaptation of height in UK Biobank . eLife , 8 , e39725 . OpenUrl CrossRef PubMed ↵ Boyce , W. T. , Sokolowski , M. B. , & Robinson , G. E. ( 2020 ). Genes and environment, development and time . Proc. Nat. Acad. Sci. USA , 117 , 23235 – 23241 . OpenUrl Abstract / FREE Full Text ↵ Broekmans , F. J. , Faddy , M. J. , Scheffer , G. , & te Velde , E. R. ( 2004 ). Antral follicle counts are related to age at natural fertility loss and age at menopause . Menopause , 11 , 607 – 614 . OpenUrl CrossRef PubMed ↵ Brunel , N. , Hakim , V. , Isope , P. , Nadal , J.-P. , & Barbour , B. ( 2004 ). Optimal information storage and the distribution of synaptic weights . Neuron , 43 , P745 – 757 . OpenUrl ↵ Buechel , S. D. , Booksmythe , I. , Kotrschal , A. , Jennions , M. D. , & Kolm , N. ( 2016 ). Artificial selection on male genitalia length alters female brain size . Proc R. Soc. B , 283 , 20161796 . OpenUrl CrossRef PubMed ↵ Burnham , K. P. , & Anderson , D. R. ( 2002 ). Model selection and multimodel inference (2nd ed .). New York, USA : Springer . ↵ Bush , S. J. , & Goriely , A. ( 2024 ). Can the male germline offer insight into mammalian brain size expansion? Andrology , 1 – 7 . ↵ Bush , S. J. , & Goriely , A. ( 2025 ). A role for the male germline in the expansion of the mammalian brain . bioRxiv . doi: 10.1101/2025.09.15.676184 OpenUrl Abstract / FREE Full Text ↵ Canela-Xandri , O. , Rawlik , K. , & Tenesa , A. ( 2018 ). An atlas of genetic associations in UK Biobank . Nat. Genet ., 50 , 1593 – 1599 . OpenUrl CrossRef PubMed ↵ Caswell , H. ( 1989 ). Analysis of life table response experiments i. decomposition of effects on population growth rate . Ecol. Model ., 46 , 221 – 237 . OpenUrl CrossRef Web of Science ↵ Caswell , H. ( 2001 ). Matrix population models (2nd ed .). Sunderland, MA, USA : Sinauer . ↵ Caswell , H. ( 2019 ). Sensitivity analysis: Matrix methods in demography and ecology . Cham, Switzerland : Springer Open . ↵ Champagnat , N. ( 2006 ). A microscopic interpretation for adaptive dynamics trait substitution sequence models . Stoch. Process. Their Appl ., 116 , 1127 – 1160 . OpenUrl ↵ Chang , M.-Y. , Chiang , C.-H. , Hsieh , T.-T. , Soong , Y.-K. , & Hsu , K.-H. ( 1998 ). Use of the antral follicle count to predict the outcome of assisted reproductive technologies . Fertil. Steril ., 69 , 505 – 510 . OpenUrl CrossRef PubMed Web of Science ↵ Coqueugniot , H. , Hublin , J.-J. , Veillon , F. , Houët , F. , & Jacob , T. ( 2004 ). Early brain growth in Homo erectus and implications for cognitive ability . Nature , 431 , 299 – 302 . OpenUrl CrossRef PubMed Web of Science ↵ Corral-López , A. , Bloch , N. I. , Kotrschal , A. , van der Bijl , W. , Buechel , S. D. , Mank , J. E. , & Kolm , N. ( 2017 ). Female brain size affects the assessment of male attractiveness during mate choice . Sci. Adv ., 3 , e1601990 . OpenUrl FREE Full Text ↵ Cox , S. , Ritchie , S. , Fawns-Ritchie , C. , Tucker-Drob , E. , & Deary , I. ( 2019 ). Structural brain imaging correlates of general intelligence in UK Biobank . Intelligence , 76 , 101376 . OpenUrl CrossRef PubMed ↵ Deaner , R. O. , Isler , K. , Burkart , J. , & van Schaik , C. ( 2007 ). Overall brain size, and not encephalization quotient, best predicts cognitive ability across nonhuman primates . Brain Behav. Evol ., 70 , 115 – 124 . OpenUrl CrossRef PubMed Web of Science ↵ Dekaban , A. S. , & Sadowsky , D. ( 1978 ). Changes in brain weights during the span of human life: Relation of brain weights to body heights and body weights . Ann. Neurol ., 4 , 345 – 356 . OpenUrl CrossRef PubMed Web of Science ↵ Dieckmann , U. , & Law , R. ( 1996 ). The dynamical theory of coevolution: a derivation from stochastic ecological processes . J. Math. Biol ., 34 ( 5 ), 579 – 612 . OpenUrl CrossRef PubMed Web of Science ↵ J.H. Kaas & S. Herculano-Houzel Drennan , A. , Sherwood , C. C. , & Gómez-Robles , A. ( 2026 ). The evolution of extended neurodevelopment and increased neuroplasticity in human brains . In J.H. Kaas & S. Herculano-Houzel (Eds.), Evolution of nervous systems (3rd ed ., Vol. 4 , p. 288 – 309 ). Academic Press . OpenUrl ↵ Dyson , F. A. ( 2004 ). A meeting with Enrico Fermi . Nature , 427 , 297 . OpenUrl CrossRef PubMed Web of Science ↵ Elks , C. E. , den Hoed , M. , Zhao , J. H. , Sharp , S. J. , Sharp , S. J. , Wareham , N. J. , … Ong , K. K. ( 2012 ). Variability in the heritability of body mass index: a systematic review and meta-regression . Front. Endocrinol ., 3 . ↵ Feldman , M. W. , & Lewontin , R. C. ( 1975 ). The heritability hang-up . Science , 190 , 1163 – 1168 . OpenUrl FREE Full Text ↵ Félix , M.-A. , & Barkoulas , M. ( 2015 ). Pervasive robustness in biological systems . Nat. Rev. Genet ., 16 , 483 – 496 . OpenUrl CrossRef PubMed ↵ Freund , J. , et al. ( 2013 ). Emergence of individuality in genetically identical mice . Science , 340 , 756 – 759 . OpenUrl Abstract / FREE Full Text ↵ Futuyma , D. J. , & Kirkpatrick , M. ( 2017 ). Evolution (4th ed .). Sunderland, MA, USA : Sinauer Associates . ↵ Gluckman , P. D. , & Hanson , M. A. ( 2006 ). Evolution, development and timing of puberty . Trends Endocrinol. Metab ., 17 , 7 – 12 . OpenUrl CrossRef PubMed Web of Science ↵ Gómez-Robles , A. , Hopkins , W. D. , Schapiro , S. J. , & Sherwood , C. C. ( 2015 ). Relaxed genetic control of cortical organization in human brains compared with chimpanzees . Proc. Natl. Acad. Sci. USA , 112 , 14799 – 14804 . OpenUrl Abstract / FREE Full Text ↵ Gómez-Robles , A. , Hopkins , W. D. , & Sherwood , C. C. ( 2013 ). Increased morphological asymmetry, evolvability and plasticity in human brain evolution . Proc. R. Soc. B ., 280 , 20130575 . OpenUrl CrossRef PubMed ↵ Gómez-Robles , A. , Nicolaou , C. , Smaers , J. B. , & Sherwood , C. C. ( 2024 ). The evolution of human altriciality and brain development in comparative context . Nat. Ecol. Evol ., 8 , 133 – 146 . OpenUrl PubMed ↵ González-Forero , M. ( 2024a ). Evolutionary-developmental (evo-devo) dynamics of hominin brain size . Nat. Hum. Behav ., 8 , 1321 – 1333 . OpenUrl PubMed ↵ González-Forero , M. ( 2024b ). A mathematical framework for evo-devo dynamics . Theor. Popul. Biol ., 155 , 24 – 50 . OpenUrl CrossRef PubMed ↵ González-Forero , M. ( 2026 ). Computer code for: Mechanistic quantification of the effects of genes, environment, and development on human behaviour . Zenodo , doi: 10.5281/zenodo.19881503 . OpenUrl CrossRef ↵ González-Forero , M. , Faulwasser , T. , & Lehmann , L. ( 2017 ). A model for brain life history evolution . PLOS Comp. Biol ., 13 , e1005380 . OpenUrl CrossRef ↵ González-Forero , M. , & Gardner , A. ( 2018 ). Inference of ecological and social drivers of human brain-size evolution . Nature , 557 , 554 – 557 . OpenUrl CrossRef PubMed ↵ González-Forero , M. , & Gómez-Robles , A. ( 2025 ). Why did the human brain size evolve? A way forward . Phil. Trans. R. Soc. B , 380 , 20240114 . OpenUrl CrossRef PubMed ↵ Gunz , P. , Neubauer , S. , Falk , D. , Tafforeau , P. , Cabec , A. L. , Smith , T. M. , … Alemseged , Z. ( 2020 ). Australopithecus afarensis endocasts suggest ape-like brain organization and prolonged brain growth . Sci. Adv ., 6 , eaaz4729 . OpenUrl FREE Full Text ↵ Gunz , P. , Neubauer , S. , Maureille , B. , & Hublin , J.-J. ( 2010 ). Brain development after birth differs between Neanderthals and modern humans . Curr. Biol ., 20 , R921 – R922 . OpenUrl CrossRef PubMed ↵ Haworth , C. M. A. , Wright , M. J. , Luciano , M. , Martin , N. G. , de Geus , E. J. C. , van Beijsterveldt , C. E. M. , … Plomin , R. ( 2010 ). The heritability of general cognitive ability increases linearly from childhood to young adulthood . Mol. Psychiatry , 15 , 1112 – 1120 . OpenUrl CrossRef PubMed Web of Science ↵ Heldstab , S. A. , Isler , K. , Schuppli , C. , & van Schaik , C. P. ( 2020 ). When ontogeny recapitulates phylogeny: Fixed neurodevelopmental sequence of manipulative skills among primates . Sci. Adv ., 6 , eabb4685 . OpenUrl FREE Full Text ↵ Heldstab , S. A. , Kosonen , Z. K. , Koski , S. E. , Burkart , J. M. , van Schaik , C. P. , & Isler , K. ( 2016 ). Manipulation complexity in primates coevolved with brain size and terrestriality . Sci. Rep ., 6 , 24528 . OpenUrl CrossRef PubMed ↵ Henshaw , J. M. , Morrissey , M. B. , & Jones , A. G. ( 2020 ). Quantifying the causal pathways contributing to natural selection . Evolution , 74 , 2560 – 2574 . OpenUrl CrossRef PubMed ↵ Howarth , C. , Peppiatt-Wildman , C. M. , & Attwell , D. ( 2010 ). The energy use associated with neural computation in the cerebellum . J. Cereb. Blood Flow Metab ., 30 , 403 – 414 . OpenUrl CrossRef PubMed Web of Science ↵ Hrvoj-Mihic , B. , Bienvenu , T. , Stefanacci , L. , Muotri , A. R. , & Semendeferi , K. ( 2013 ). Evolution, development, and plasticity of the human brain: from molecules to bones . Front. Hum. Neurosci ., 7 , 707 . OpenUrl CrossRef PubMed Jiska S. Peper , R.M.B. , Boomsma , D. I. , Kahn , R. S. , & Pol , H. E. H. ( 2007 ). Genetic influences on human brain structure: A review of brain imaging studies in twins . Hum. Brain Mapp ., 28 , 464 – 473 . OpenUrl CrossRef PubMed Web of Science ↵ Johnson , W. ( 2007 ). Genetic and environmental influences on behavior: capturing all the interplay . Psychol. Rev ., 114 , 423 – 440 . OpenUrl CrossRef PubMed Web of Science ↵ Kaplan , H. , Hill , K. , Lancaster , J. , & Hurtado , A. M. ( 2000 ). A theory of human life history evolution: diet, intelligence, and longevity . Evol. Anthropol ., 9 , 156 – 185 . OpenUrl CrossRef Web of Science ↵ Kievit , R. A. , et al. ( 2018 ). Developmental cognitive neuroscience using latent change score models: A tutorial and applications . Dev. Cogn. Neurosci ., 33 , 99 – 117 . OpenUrl CrossRef PubMed ↵ Kitano , H. ( 2007 ). Towards a theory of biological robustness . Mol. Sys. Biol ., 3 , 137 . OpenUrl FREE Full Text ↵ Koster , J. , et al. ( 2020 ). The life history of human foraging: Cross-cultural and individual variation . Sci. Adv ., 6 , eaax9070 . OpenUrl FREE Full Text ↵ Kotrschal , A. , Corral-Lopez , A. , Zajitschek , S. , Immler , S. , Maklakov , A. A. , & Kolm , N. ( 2015 ). Positive genetic correlation between brain size and sexual traits in male guppies artificially selected for brain size . J. Evol. Biol ., 28 , 841 – 850 . OpenUrl CrossRef PubMed ↵ Lehtonen , J. , & Otsuka , J. ( 2023 ). Evolutionary game theory of continuous traits from a causal perspective . Phil. Trans. R. Soc. B , 378 , 20210507 . OpenUrl CrossRef PubMed ↵ León , J. A. ( 1976 ). Life histories as adaptive strategies . J. Theor. Biol ., 60 , 301 – 335 . OpenUrl CrossRef PubMed Web of Science ↵ MacLean , E. L. , Hare , B. , Nunn , C. L. , Addessi , E. , Amici , F. , Anderson , R. C. , … Zhao , Y. ( 2014 ). The evolution of self-control . Proc. Natl. Acad. Sci. USA , 111 , E2140E2148 . OpenUrl ↵ Matos , B. , Publicover , S. J. , Castro , L. F. C. , Esteves , P. J. , & Fardilha , M. ( 2021 ). Brain and testis: more alike than previously thought? Open Biol ., 11 , 200322 . OpenUrl CrossRef PubMed ↵ McElreath , R. ( 2018 ). Sizing up human brain evolution . Nature , 557 , 496 – 497 . OpenUrl PubMed ↵ McElreath , R. ( 2020 ). Statistical rethinking (2nd ed .). Boca Raton, FL, USA : CRC Press . ↵ McHenry , H. M. , & Coffing , K. ( 2000 ). Australopithecus to Homo: transformations in body and mind . Annu. Rev. Anthropol ., 29 , 125 – 146 . OpenUrl CrossRef Web of Science ↵ S. van Strien & S. V. Lunel Metz , J. , Geritz , S. , Meszéna , G. , Jacobs , F. , & van Heerwaarden , J. ( 1996 ). Adaptive dynamics, a geometrical study of the consequences of nearly faithful reproduction . In S. van Strien & S. V. Lunel (Eds.), Stochastic and spatial structures of dynamical systems (p. 183 – 231 ). Amsterdam, Netherlands : Konink. Nederl. Akad. Wetensch. Verh. Afd. Natuurk. Eerste Reeks . ↵ Miller , D. J. , Duka , T. , Stimpson , C. D. , Schapiro , S. J. , Baze , W. B. , McArthur , M. J. , … Sherwood , C. C. ( 2012 ). Prolonged myelination in human neocortical evolution . Proc. Natl. Acad. Sci. USA , 109 , 16480 – 16485 . OpenUrl Abstract / FREE Full Text ↵ Mitchell , K. J. ( 2018 ). Innate: How the wiring of our brains shapes who we are . Princeton, NJ, USA : Princeton Univ. Press . ↵ Molz , B. , Eising , E. , Alagöz , G. , Schijven , D. , Francks , C. , Gunz , P. , & Fisher , S. E. ( 2024 ). Imaging genomics reveals genetic architecture of the globular human braincase . bioRxiv . doi: 10.1101/2024.03.20.585712 OpenUrl Abstract / FREE Full Text ↵ Nave , G. , Jung , W. H. , Linnér , R. K. , Kable , J. W. , & Koellinger , P. D. ( 2019 ). Are bigger brains smarter? Evidence from a large-scale preregistered study . Psychol. Sci ., 30 , 43 – 54 . OpenUrl CrossRef PubMed ↵ Ota , M. , Spence , J. P. , Zeng , T. , Dann , E. , Marson , A. , & Pritchard , J. K. ( 2025 ). Causal modeling of gene effects from regulators to programs to traits: integration of genetic associations and perturb-seq . bioRxiv , doi: 10.1101/2025.01.22.634424 . OpenUrl Abstract / FREE Full Text ↵ Paige , J. , & Perreault , C. ( 2024 ). 3.3 million years of stone tool complexity suggests that cumulative culture began during the Middle Pleistocene . Proc. Natl. Acad. Sci ., 121 , e2319175121 . OpenUrl CrossRef PubMed ↵ Pearl , J. ( 2009 ). Causality (2nd ed .). Cambridge Univ. Press . ↵ Plomin , R. , & Kawakami , K. ( 2025 ). Nurture and nonshared environment in cognitive development . Psychol. Rev .. doi: 10.1037/rev0000551 OpenUrl CrossRef ↵ Robson , S. L. , & Wood , B. ( 2008 ). Hominin life history: reconstruction and evolution . J. Anat ., 212 , 394 – 425 . OpenUrl CrossRef PubMed Web of Science ↵ Roebroeks , W. , & Villa , P. ( 2011 ). On the earliest evidence for habitual use of fire in Europe . Proc. Natl. Acad. Sci ., 108 , 5209 – 5214 . OpenUrl Abstract / FREE Full Text ↵ Rosenberg , K. R. ( 2021 ). The evolution of human infancy: why it helps to be helpless . Ann. Rev. Anthropol ., 50 , 423 – 440 . OpenUrl CrossRef ↵ Rosenberg , M. D. , Casey , B. J. , & Holmes , A. J. ( 2018 ). Prediction complements explanation in understanding the developing brain . Nat. Commun ., 9 , 589 . OpenUrl CrossRef PubMed ↵ Sachser , N. , Hennessy , M. B. , & Kaiser , S. ( 2018 ). The adaptive shaping of social behavioural phenotypes during adolescence . Biol. Lett ., 14 , 20180536 . OpenUrl CrossRef PubMed ↵ Sachser , N. , Zimmermann , T. D. , Hennessy , M. B. , & Kaiser , S. ( 2020 ). Sensitive phases in the development of rodent social behavior . Curr. Opin. Behav. Sci ., 36 , 63 – 70 . OpenUrl PubMed ↵ Sakaue , S. , et al. ( 2021 ). A cross-population atlas of genetic associations for 220 human phenotypes . Nat. Genet ., 53 , 1415 – 1424 . OpenUrl CrossRef PubMed ↵ Salazar-Ciudad , I. , & Jernvall , J. ( 2010 ). A computational model of teeth and the developmental origins of morphological variation . Nature , 464 , 583 – 586 . OpenUrl CrossRef PubMed Web of Science ↵ Schuppli , C. , Isler , K. , & van Schaik , C. P. ( 2012 ). How to explain the unusually late age at skill competence among humans . J. Hum. Evol ., 63 , 843 – 850 . OpenUrl CrossRef PubMed Web of Science ↵ Sheth , R. , et al. ( 2012 ). Hox genes regulate digit patterning by controlling the wavelength of a Turing-type mechanism . Science , 338 , 1476 – 1480 . OpenUrl Abstract / FREE Full Text ↵ Snell-Rood , E. C. ( 2013 ). An overview of the evolutionary causes and consequences of behavioural plasticity . Anim. Behav ., 85 , 1004 – 1011 . OpenUrl CrossRef Web of Science ↵ Sohail , M. , et al. ( 2019 ). Polygenic adaptation on height is overestimated due to uncorrected stratification in genome-wide association studies . eLife , 8 , e39702 . OpenUrl CrossRef PubMed ↵ Sotiras , A. , Toledo , J. B. , Gur , R. E. , Gur , R. C. , Satterthwaite , T. D. , & Davatzikos , C. ( 2017 ). Patterns of coordinated cortical remodeling during adolescence and their associations with functional specialization and evolutionary expansion . Proc. Natl. Acad. Sci. USA , 114 , 3527 – 3532 . OpenUrl Abstract / FREE Full Text ↵ Sparling , J. , Dragomir , C. , Ramey , S. , & Florescu , L. ( 2005 ). An educational intervention improves developmental progress of young children in a Romanian orphanage . Infant Ment. Health J ., 26 , 127 – 142 . OpenUrl CrossRef PubMed Web of Science ↵ Steuber , V. , Mittmann , W. , Hoebeek , F. E. , Silver , R. A. , De Zeeuw , C. I. , Häusser , M. , & De Schutter , E. ( 2007 ). Cerebellar LTD and pattern recognition by Purkinje cells . Neuron , 54 , P121 – 136 . OpenUrl ↵ Tan , T. , et al. ( 2024 ). Family-GWAS reveals effects of environment and mating on genetic associations . medRxiv , doi: 10.1101/2024.10.01.24314703 . OpenUrl Abstract / FREE Full Text ↵ Turkheimer , E. , & Waldron , M. ( 2000 ). Nonshared environment: A theoretical, methodological, and quantitative review . Psychol. Bull ., 126 , 78 – 108 . OpenUrl CrossRef PubMed Web of Science ↵ van Schaik , C. P. , & Burkart , J. M. ( 2011 ). Social learning and evolution: the cultural intelligence hypothesis . Phil. Trans. R. Soc. B , 366 , 1008 – 1016 . OpenUrl CrossRef PubMed ↵ Walasek , N. , Panchanathan , K. , & Frankenhuis , W. E. ( 2025 ). The evolution of sensitive periods beyond early ontogeny: Bridging theory and data . Funct. Ecol ., 39 , 2962 – 2973 . OpenUrl ↵ West , G. B. , Brown , J. H. , & Enquist , B. J. ( 2001 ). A general model for ontogenetic growth . Nature , 413 , 628 – 631 . OpenUrl CrossRef GeoRef PubMed Web of Science ↵ Zeng , H. , Shen , E. H. , Hohmann , J. G. , Oh , S. W. , Bernard , A. , Royall , J. J. , … Jones , A. R. ( 2012 ). Large-scale cellular-resolution gene profiling in human neocortex reveals species-specific molecular signatures . Cell , 149 , 483 – 496 . OpenUrl CrossRef PubMed Web of Science View the discussion thread. Back to top Previous Next Posted May 05, 2026. Download PDF Supplementary Material Data/Code Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following Mechanistic quantification of the effects of genes, environment, and development on human behaviour Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share Mechanistic quantification of the effects of genes, environment, and development on human behaviour Mauricio González-Forero , Aida Gómez-Robles bioRxiv 2025.03.17.643651; doi: https://doi.org/10.1101/2025.03.17.643651 Share This Article: Copy Citation Tools Mechanistic quantification of the effects of genes, environment, and development on human behaviour Mauricio González-Forero , Aida Gómez-Robles bioRxiv 2025.03.17.643651; doi: https://doi.org/10.1101/2025.03.17.643651 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Evolutionary Biology Subject Areas All Articles Animal Behavior and Cognition (7635) Biochemistry (17697) Bioengineering (13895) Bioinformatics (41951) Biophysics (21456) Cancer Biology (18594) Cell Biology (25520) Clinical Trials (138) Developmental Biology (13381) Ecology (19903) Epidemiology (2067) Evolutionary Biology (24323) Genetics (15612) Genomics (22510) Immunology (17738) Microbiology (40401) Molecular Biology (17184) Neuroscience (88622) Paleontology (667) Pathology (2833) Pharmacology and Toxicology (4825) Physiology (7644) Plant Biology (15158) Scientific Communication and Education (2046) Synthetic Biology (4296) Systems Biology (9825) Zoology (2271)","source_license":"CC-BY-4.0","license_restricted":false}