Searches
/
Philosophical Transactions Of The Royal Society Of London. Series B, Biological Sciences[JOURNAL]
Philosophical Transactions Of The Royal Society Of London. Series B, Biological Sciences[JOURNAL]
Sun
200 papers
Essential is essential: a review evaluating the linkage of essential amino acid deficiency resulting from poor access to animal source foods with environmental enteric dysfunction and stunting.
Khan S, Margolis K, Wallach T
Philos Trans R Soc Lond B Biol Sci · 2026 May · PMID 42132024 · Publisher ↗
Addressing endemic stunting remains a primary United Nations development goal. One potential contributor to global linear growth failure is environmental enteric dysfunction (EED). However, it has proven challenging to a...
Addressing endemic stunting remains a primary United Nations development goal. One potential contributor to global linear growth failure is environmental enteric dysfunction (EED). However, it has proven challenging to address with interventions focused on water, sanitation and hygiene. Initial theories of EED placed primacy on recurrent enteric exposures; however, it appears that these exposures are inadequate to explain the phenotype and sequelae. Children in EED endemic regions are at high risk of having inadequate nutrition, and strong associations can be found between malnutrition and stunting. In this review, we summarize the clinical, translational and mechanistic evidence linking intake of animal source foods as a source of protein and micro-nutrients and, in their absence, essential amino acid (EAA) deficiency with stunted growth and features of EED such as altered immune behaviour. EAA deficiency has been linked to growth failure through several mechanistic pathways including intestinal inflammation, barrier disruption and chondral plate closure, and amino acid supplementation has shown clinical efficacy. By better understanding this linkage, we will be able to better study not only EED but also pathways in which dietary sources mechanistically alter systemic signalling in metabolism and immune function. This article is part of the theme issue 'Biological, biomedical and environmental drivers of stunting'.
Dismantling silos: what we have learnt about the relationship between child wasting and stunting, and the implications for connected policy, research and programmes.
Khara T, Black R, Briend A … +1 more , Thurstans S
Philos Trans R Soc Lond B Biol Sci · 2026 May · PMID 42132023 · Publisher ↗
Since 2014, the Wasting and Stunting Technical Interest Group and others have amassed a body of evidence on the relationship between child wasting and stunting. The resulting research papers and reports have enhanced our...
Since 2014, the Wasting and Stunting Technical Interest Group and others have amassed a body of evidence on the relationship between child wasting and stunting. The resulting research papers and reports have enhanced our understanding of undernutrition, particularly how children become wasted and/or stunted and how they experience these conditions and their consequences. Evidence highlights the common determinants and interconnected physiological processes leading to a child becoming wasted and/or stunted and we have built a greater understanding of how wasting treatment, or its lack, impacts children's linear growth. We also know more about the mortality-risk consequences for a child being concurrently wasted and stunted, how commonly this occurs, which children are particularly at risk and how to best identify and potentially support them. Based on this evidence, some shifts to policy and programme approaches, and to how research is conducted, are indicated in order to more effectively address both forms of undernutrition. These include conducting common contextual causal analysis, developing common prevention strategies that target overlapping drivers including in utero, considering wasting as part of the pathway to stunting, and vice versa when designing prevention strategies and including children at greatest mortality risk owing to multiple deficits in treatment targeting. This article is part of the theme issue 'Biological, biomedical and environmental drivers of stunting'.
Characterizing viral clearance kinetics in acute influenza.
Wongnak P, Seers T, Jittamala P … +4 more , Imwong M, Schilling W, Watson J, White NJ
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057729 · Publisher ↗
Pharmacometric assessment of antiviral efficacy in acute influenza informs treatment decisions and pandemic preparedness. We characterized natural viral clearance in acute influenza to guide phase II trial design using s...
Pharmacometric assessment of antiviral efficacy in acute influenza informs treatment decisions and pandemic preparedness. We characterized natural viral clearance in acute influenza to guide phase II trial design using simulations based upon observed data. Standardized duplicate oropharyngeal swabs were collected daily over 14 days from 80 untreated low-risk Thai adults, with viral densities measured using quantitative polymerase chain reaction. We evaluated three models to describe viral clearance: exponential, bi-exponential and growth-and-decay. The growth-and-decay model provided the best fit, but the exponential decay model was the most parsimonious. The median viral clearance half-life was 10.3 h (interquartile range (IQR): 6.8-15.4h), varying by influenza type: 9.6 h (IQR: 6.2-13.0 h) for influenza A and 14.0 h (IQR: 10.3-19.3 h) for influenza B. Simulated trials using parameters from the exponential decay model showed that 148 patients per arm provide over 90% power to detect treatments accelerating viral clearance by 40%. Variation in clearance rates strongly impacted the power; doubling this variation would require 232 patients per arm for an antiviral with a 60% effect size. A sampling strategy with four swabs per day reduces the required sample size to 81 per arm while maintaining over 80% power. We recommend this approach to assess and compare current anti-influenza drugs. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
Blood viral density dynamics in acute Chikungunya virus infection.
Souza-Lima EC, Souza RP, Dutra JVR … +5 more , Ascencao FR, Schilling WHK, White NJ, Aguiar RS, Teixeira MM
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057728 · Publisher ↗
Chikungunya infection is caused by the Chikungunya alpha virus. The disease is characterized by an acute febrile illness that evolves to chronic, often severe, polyarticular limb arthralgia in approximately 30% of patien...
Chikungunya infection is caused by the Chikungunya alpha virus. The disease is characterized by an acute febrile illness that evolves to chronic, often severe, polyarticular limb arthralgia in approximately 30% of patients. There are no specific treatments for Chikungunya. To characterize the dynamics of viraemia in order to plan future antiviral studies, we took daily blood samples and evaluated serial viral densities measured by RT-PCR in a cohort of 19 adult Brazilian patients with acute Chikungunya infection. Viraemia exhibited a rapid decline, and most individuals were RT-PCR negative by day 4 after study entry. Higher viral loads were associated with earlier clinical presentations. Concentrations of sTNFR1 closely followed the decline in Chikungunya viral loads. Viral clearance measurements may be useful for the development and evaluation of novel treatments for Chikungunya infection. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
The application of treatment response biomarkers in clinical trials for cryptococcosis, invasive aspergillosis and candidiasis.
Allebone-Salt P, Wake R, White PL … +3 more , Price J, Harrison T, Bicanic T
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057727 · Publisher ↗
Just three classes of antifungals are available for the treatment of WHO critical priority fungal pathogens Cryptococcus, Candida and Aspergillus spp. Drug development and optimization via phase II and III clinical trial...
Just three classes of antifungals are available for the treatment of WHO critical priority fungal pathogens Cryptococcus, Candida and Aspergillus spp. Drug development and optimization via phase II and III clinical trials is reliant on binary composite endpoints (clinical/radiological/mycological), requiring large patient cohorts with challenging logistics at great expense. The validation and implementation of continuous mycological treatment response biomarkers have the potential to facilitate smaller investigator-led trials to optimize existing and novel drugs and combinations, as demonstrated in cryptococcal meningitis, where early fungicidal activity is now a primary endpoint in phase II trials and a secondary endpoint in phase III trials. In invasive aspergillosis, serum galactomannan holds promise: numerous studies show a decline on treatment to be associated with clinical outcome, however, further prospective, standardized studies are required before inclusion in clinical trial response criteria. In invasive candidiasis, while beta-D-glucan (BDG) and qPCR in invasive candidiasis decline during treatment, studies are limited to inadequately powered retrospective cohorts with simple correlation analyses. We include preliminary data from our ongoing prospective study showing an overall decline on treatment, with qPCR clearing more rapidly than BDG, with considerable heterogeneity of kinetics in this complex population. Future analyses must account for baseline values and inter-patient variation. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
Use of animal efficacy, pharmacokinetic/pharmacodynamic and first in human data for regulatory approval of treatment options against chemical, biological, radiological and nuclear threats.
Buchholz S, Cavaleri M
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057726 · Publisher ↗
Chemical, biological, radiological and nuclear (CBRN) agents can pose a significant risk to public health. Concerningly, for many of the CBRN threats there are no treatment options available. Conducting pivotal randomize...
Chemical, biological, radiological and nuclear (CBRN) agents can pose a significant risk to public health. Concerningly, for many of the CBRN threats there are no treatment options available. Conducting pivotal randomized controlled trials (RCTs) usually required for regulatory approval of medicinal products may not be feasible for medical countermeasures (MCM) intended for the treatment of CBRN threats because of the low prevalence and case numbers and ethical concerns on conducting RCTs, considering the potential serious health outcomes associated with CBRN threats. This is concerning, as it hampers the development and availability of new treatments against CBRN threats, which are needed for preparedness to ensure appropriate response to an outbreak. This review outlines alternative, regulatory flexible approaches that were used for the approval under exceptional circumstances of therapeutic MCMs against biological threats in the European Union. Considerations on the requirements of using animal efficacy data as key evidence for inferring efficacy from animals to humans, on using animal pharmacokinetic/pharmacodynamic data including pharmacometric modelling and simulation approaches to predict the human dose and on generating safety data in healthy humans are discussed. Challenges and advantages of this approach are highlighted, including the lessons learned based on the examples of tecovirimat, zanamivir and obiltoxaximab. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
Exposure history shapes SARS-CoV-2 viral dynamics in non-human primates and provides insights into correlates of protection against infection and transmission.
Mitard de Girardier A, Hérate C, Marlin R … +12 more , Donati F, Rahou Y, Bossevot L, Sconosciuti Q, Cavarelli M, Dereuddre-Bosquet N, Relouzat F, Van der Werf S, Simon-Loriere E, Prague M, Guedj J, Le Grand R
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057725 · Publisher ↗
In the SARS-CoV-2 Omicron-variant era, the efficacy of vaccination in preventing viral replication is poorly understood and is further complicated by heterogeneous and hybrid immunity in the population. Here, we analysed...
In the SARS-CoV-2 Omicron-variant era, the efficacy of vaccination in preventing viral replication is poorly understood and is further complicated by heterogeneous and hybrid immunity in the population. Here, we analysed by mathematical models the kinetics of both viral load and antibodies in non-human primates with different exposure histories, conferred by infection or vaccination with mono- or bi-valent vaccines. The model describes the control of viral replication by antibodies that accelerate the clearance and reduce the infectivity of viral particles, and suggests that a previous infection leads to a faster elimination of infected cells, which is not mediated by antibodies. Finally, we used the model to simulate natural infections and show that binding, neutralization and exposure history all need to be taken into account to predict risks of infection and transmission. Protection from infection could be reached at binding levels of 3 × 104 AU ml-1 for individuals who received both monovalent and bivalent vaccinations, whereas hybrid immunity (infection followed by bivalent vaccination) could provide complete protection irrespective of antibody levels. These findings highlight the importance of exposure history in shaping immune responses and suggest correlates of protection for future vaccine and monoclonal antibodies strategies. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
Integrating viral kinetics into routine outbreak surveillance: challenges, opportunities and lessons from COVID-19.
Hay JA, Larremore DB, Aatresh AV … +1 more , Kissler S
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057724 · Publisher ↗
Viral kinetics provide crucial insights into the biology and epidemiology of infections, with direct implications for basic science, therapeutics development and policy. The COVID-19 pandemic showcased the power of viral...
Viral kinetics provide crucial insights into the biology and epidemiology of infections, with direct implications for basic science, therapeutics development and policy. The COVID-19 pandemic showcased the power of viral kinetics surveillance and modelling; however, our understanding of viral kinetics has been limited to retrospective analyses, convenience samples and bespoke models. To strengthen responses to ongoing and emerging outbreaks, we argue that viral kinetics should be a core component of pathogen surveillance. Building upon insights gained during the COVID-19 pandemic, we review ways that continuous viral kinetic surveillance supports infectious disease response by informing epidemiological parameters, development and deployment of therapeutics, and adaptive policy design. To achieve this, various challenges must be addressed regarding data standards, study design and communication. We advocate for the creation of a global, living library of viral kinetics data, with associated data sharing standards, modelling toolkits and on-demand epidemiological reports. Successfully integrating viral kinetics into active disease surveillance efforts will support both active outbreak response and improve the knowledge base vital for pandemic preparedness. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
Dengue virus non-structural protein 1 as a surrogate biomarker of antiviral efficacy in early phase clinical trials.
Metem V, Mairiang D, Ding J … +16 more , Thiemmeca S, Songjaeng A, Kongmanas K, Punyadee N, Poraha R, Sayboonruan P, Prommool T, Prayongkul D, Tangthawornchaikul N, Puttikhunt C, Limpitikul W, Kittirat P, Sriruksa K, Malasit P, Tarning J, Avirutnan P
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057723 · Publisher ↗
Dengue virus (DENV) infection imposes a major global health burden, with around 100 million symptomatic cases and about 40 000 deaths each year. Despite decades of effort, no direct-acting antivirals are licensed and cur...
Dengue virus (DENV) infection imposes a major global health burden, with around 100 million symptomatic cases and about 40 000 deaths each year. Despite decades of effort, no direct-acting antivirals are licensed and current vaccines show serotype-dependent and baseline serostatus-dependent efficacy, raising concerns about antibody-dependent enhancement. Early phase therapeutic trials therefore rely mainly on quantitative viraemia as a virological endpoint. However, DENV RNA in blood becomes undetectable soon after presentation, and may be an imperfect surrogate for infectious virus owing to defective interfering particles and continued replication in tissues. The secreted non-structural protein 1 (NS1) correlates with viral replication and disease severity in many studies and remains detectable after virological clearance, making it an attractive complementary biomarker. Here, we review mechanistic and clinical evidence linking NS1 to viral burden, vascular leakage and thrombocytopenia, and summarize data from human challenge studies, natural infection cohorts and pharmacometric modelling. We highlight the extended analytical window and growing availability of quantitative NS1 assays and propose practical analysis frameworks for incorporating NS1 into early phase dengue trials alongside viral RNA. We also discuss limitations, including serotype, immune status and assay-dependent effects and outline priorities for standardization of NS1-based endpoints. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
Antiviral treatment for influenza.
Hui DS
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057722 · Publisher ↗
Seasonal influenza is a widespread acute respiratory infection that causes significant illness and death worldwide. Two major antiviral classes are neuraminidase inhibitors (NAIs) and polymerase inhibitors. NAIs, includi...
Seasonal influenza is a widespread acute respiratory infection that causes significant illness and death worldwide. Two major antiviral classes are neuraminidase inhibitors (NAIs) and polymerase inhibitors. NAIs, including oseltamivir, zanamivir, peramivir and laninamivir, block viral release, while polymerase inhibitors such as baloxavir disrupt viral RNA replication. Early administration within 48 h of symptom onset reduces illness duration, severity and complications, particularly in high-risk groups. Oseltamivir is the most widely studied NAI, demonstrating reduced viral shedding, faster symptom resolution and lower complication rates, though gastrointestinal side effects are common. Higher doses generally do not improve outcomes compared to standard dosing. Zanamivir is more effective against influenza B and is inhibitory for most influenza A viruses resistant to oseltamivir, but the inhaled formulation is less suitable for patients with severe illness or airway disease. Intravenous (IV) zanamivir is approved for hospitalized influenza patients in some countries. Peramivir offers IV treatment options, while laninamivir is mainly used in Japan. Baloxavir shows superior viral load reduction and comparable symptom relief to oseltamivir in outpatients, though resistance variants can emerge. Favipiravir and newer polymerase inhibitors are under investigation. Combination therapies may enhance recovery, with limited evidence. Overall, timely antiviral use is critical to reducing influenza's burden. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
How pharmacokinetic/pharmacodynamic models can inform the development of new HIV medications.
Miranda SE, Khoo S
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057721 · Publisher ↗
There remains a need for new HIV medications. This has been challenged by the increasing costs associated with drug discovery and testing. One solution to this issue involves the incorporation of pharmacokinetic/pharmaco...
There remains a need for new HIV medications. This has been challenged by the increasing costs associated with drug discovery and testing. One solution to this issue involves the incorporation of pharmacokinetic/pharmacodynamic (PK/PD) models into the development pipeline. This article provides an overview of how PK/PD models can be used to provide insights to aid this endeavour. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
Bacterial quantitation in clinical infection: a systematic review.
Benjamin A, Phommasone K, White NJ … +1 more , Ashley EA
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057720 · Publisher ↗
Pivotal clinical studies of new anti-infective drugs enrol patients with strictly defined clinical syndromes, with clinical and laboratory endpoints used to define treatment success. Assessing bacterial killing in the ho...
Pivotal clinical studies of new anti-infective drugs enrol patients with strictly defined clinical syndromes, with clinical and laboratory endpoints used to define treatment success. Assessing bacterial killing in the host by serial quantitation has potential as a more efficient pharmacometric approach to assess antibiotic efficacy. We conducted a systematic review of the use of bacterial quantitation in clinical infection. The main syndromes studied in the 222 included articles were bacteraemia, chronic respiratory disease, diarrhoea, pneumonia and sexually transmitted infections. Most reports (135 articles, 61%) quantified bacterial load in different specimen types using quantitative polymerase chain reaction (qPCR). For bacteraemias, bacterial DNA load measured in whole blood by qPCR at clinical presentation was typically 2-3 log10 copies ml-1, usually substantially higher than the estimates from quantitative culture. Higher bacterial loads were associated with increased mortality in 28 of 35 studies. Faster bacterial clearance was correlated with appropriate antibiotic therapy and improved outcome in the majority of studies. Most studies sampled too infrequently for accurate characterization of bacterial clearance rates. The rate of bacterial clearance from blood or other compartments is an informative pharmacodynamic endpoint in the assessment of antibacterial therapeutic effects, but standardized approaches to assessment are needed based on optimal study design. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
Assessment of the pharmacodynamic properties of antimalarial drugs.
White NJ
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057719 · Publisher ↗
Quantifying therapeutic responses in clinical malaria is easier than for most other infections as the intraerythrocytic parasites can be counted by microscopy or estimated using quantitative PCR. In treating the blood-st...
Quantifying therapeutic responses in clinical malaria is easier than for most other infections as the intraerythrocytic parasites can be counted by microscopy or estimated using quantitative PCR. In treating the blood-stage of malaria, between 107 and 1013 parasites undergo a first-order decline in densities at a rate determined by the concentrations and potency of the antimalarial drug. A simple conceptual framework based on total intravascular parasite biomass and standard sigmoid concentration-effect relationships for parasite killing explains most, but not all, aspects of antimalarial therapeutic responses, and it has proved very useful in designing chemoprevention and treatment regimens and in understanding the selection and spread of resistance. Drugs acting on younger circulating ring-stage asexual parasites (artemisinins, cipargamin, ganaplacide) provide rapid parasite clearance, which translates into faster clinical recoveries and a life-saving benefit in severe malaria. Artemisinin-sensitive Plasmodium falciparum densities decline with a half-life (PC1/2) of usually less than 5 h. Many antimalarial drugs are eliminated slowly and provide protracted exposures, which allows full treatment to be administered in 3 days, and also provides chemosuppression of newly acquired infections for 1 month. Greater availability of drug measurement in malaria-endemic areas would facilitate the field assessment of antimalarial drugs. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
Antimicrobial pharmacodynamics: defining regimens of antibiotics that are safe, effective and prevent the emergence of resistance.
Dubey V, Reza N, Hope W
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057718 · Publisher ↗
The substantial value of modern antibacterial agents is eroded by the inexorable development of antimicrobial resistance (AMR). The discovery and development of antibiotics are central aspects of addressing AMR. Since ne...
The substantial value of modern antibacterial agents is eroded by the inexorable development of antimicrobial resistance (AMR). The discovery and development of antibiotics are central aspects of addressing AMR. Since new drug-resistant mechanisms are continually arising, ways to streamline and accelerate the development of new agents are required. An understanding of dose-exposure-response (D-E-R) relationships (also pharmacokinetic-pharmacodynamic (PK-PD)) provides a means to establish causality, i.e. demonstrating that the new antibiotic acts in a predictable way and ensuring that understanding can be harnessed for therapeutic benefit. D-E-R potentially enables the adequacy of proposed regimens for effectiveness trials to be justified, minimization or replacement of dose-finding studies, selection of regimens in special populations and selection of regimens that minimize the emergence of resistance. Here, we consider the advantages of combining empirical approaches characteristic of modern PK-PD with a mechanistic understanding of drug effect and the emergence of resistance. Conceiving the approach to new dose/schedule selection in this way provides the best chance of designing regimens that are safe and effective, and that result in resistance counterselection. In terms of the latter, regimen intensification, appropriate sequencing of antibiotics and antimicrobial combinations are relevant. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
Translational bacterial biomarkers and pharmacodynamic models in the treatment of tuberculosis.
Davies GR
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 42057717 · Publisher ↗
Treatment of Mycobacterium tuberculosis is uniquely prolonged and complex among bacterial infections due to its unusual propensity for persistence during drug stress. Persisters are subpopulations of organisms tolerant t...
Treatment of Mycobacterium tuberculosis is uniquely prolonged and complex among bacterial infections due to its unusual propensity for persistence during drug stress. Persisters are subpopulations of organisms tolerant to bactericidal antibiotics and characterized by a slower rate of elimination in the absence of genetic changes associated with resistance. The biological basis of this response is increasingly well understood and novel translational tools have become available that are capable of identifying and characterizing such organisms in vitro and in vivo. Recent in vivo and clinical observations have reaffirmed the existence of multiple subpopulations of organisms during treatment of tuberculosis (TB) but operationalizing this concept in pharmacodynamic modelling and clinical trials has been hampered by the availability of culture-based biomarkers alone. A phenomenological framework, which makes explicit use of the information from new tools, would assist with comparisons between preclinical and clinical data and may be able to support models that quantify the likely size of the pool of persister organisms from which relapses arise, facilitating more accurate predictions of the duration of therapy. Such an approach has the potential for significant translational impact in the development of new treatment regimens for TB. This article is part of the Theo Murphy meeting issue 'Evaluating anti-infective drugs'.
Global brains: the science and practice of collective intelligence at the level of whole systems.
Mulgan G
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 41987731 · Publisher ↗
What is the science and practice of collective intelligence at the level of whole systems and at a global scale? How have attempts to organize intelligence at these scales evolved, and how might they evolve in the future...
What is the science and practice of collective intelligence at the level of whole systems and at a global scale? How have attempts to organize intelligence at these scales evolved, and how might they evolve in the future? Recent research has attempted to find universal patterns in the organization of intelligence at multiple scales, from atoms and cells to birds and other animals and whole societies. This work offers insights for the design of intelligence at global scales that can support governance, whether for pandemics or climate change, and that can integrate data, models, algorithms, evidence, tacit knowledge, foresight and innovation. This opinion piece argues for analysing common features of intelligence at multiple scales; it describes some important examples of 'intelligence assemblies' in technologies, cities, professions and global initiatives; and it advocates drawing on this framework for understanding intelligence at multiple scales to accelerate both research and design for large-scale intelligence. Questions of macro-cognition design are arguably as important as micro-cognition in terms of potential social impact but they have been less comprehensively researched, partly because they cut across so many disciplinary boundaries. However, a better understanding of these phenomena may be vital for tackling the big global challenges of the twenty-first century and for humanity's future evolution. This article is part of the theme issue 'The evolution of collective intelligence'.
Collective intelligence through aggregation.
Dietrich F, List C
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 41987730 · Publisher ↗
Suppose a committee, expert panel or other group is making judgements on some issues, where these may be not just yes/no questions, such as whether a defendant is guilty, but also variables with many possible values, suc...
Suppose a committee, expert panel or other group is making judgements on some issues, where these may be not just yes/no questions, such as whether a defendant is guilty, but also variables with many possible values, such as macroeconomic or meteorological variables or travel directions. Furthermore, there may be interconnections between different issues, as in the case of economic or climate variables. How can the group arrive at 'intelligent' collective judgements, based on the group members' individual judgements? We investigate three challenges raised by this judgement-aggregation problem. First, reasonable methods of aggregation (such as defining the collective judgement for each issue as the average or median judgement) can produce inconsistent collective judgements. Second, many methods of aggregation are manipulable by strategic voting. Finally, not all methods of aggregation are conducive to tracking the truth on the issues in question. We prove new impossibility or possibility theorems on all three challenges, identifying what it takes to produce collective judgements in a consistent, non-manipulable and truth-tracking manner and thereby to achieve collective intelligence through aggregation. Overall, the median method, though imperfect, performs reasonably well. We also note the relevance of our analysis for non-human group decisions. This article is part of the theme issue 'The evolution of collective intelligence'.
Modelling the emergence of open-ended cultural evolution.
Winters J, Charbonneau M
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 41987729 · Publisher ↗
Humans stand alone in terms of their potential to collectively and cumulatively change their culture in an open-ended manner. This open-endedness provides societies with the ability to continually expand their resources...
Humans stand alone in terms of their potential to collectively and cumulatively change their culture in an open-ended manner. This open-endedness provides societies with the ability to continually expand their resources and to increase their capacity to store, transmit and process information at a collective level. Here, we propose that the production of resources arises from the interaction between cultural systems (a society's repertoire of interdependent techniques, artefacts, norms and knowledge) and search spaces (an ensemble of needs, problems and goals facing a society). Starting from this premise, we develop a macro-level model wherein both cultural systems and search spaces are subject to evolutionary dynamics. By manipulating the extent to which these dynamics are characterized by stochastic or selection-like processes, we demonstrate that open-ended growth is extremely rare, historically contingent and only possible when cultural systems and search spaces co-evolve. Here, stochastic factors must be strong enough to continually perturb the dynamics into a far-from-equilibrium state, whereas selection-like factors help maintain effectiveness and ensure the sustained production of resources. Only when this co-evolutionary dynamic maintains effective cultural systems, supports the ongoing expansion of the search space and leads to an increased provision of resources do we observe open-ended cultural evolution. This article is part of the theme issue 'The evolution of collective intelligence'.
Crowdsourced versus large language models forecasting: evidence for the accuracy-correlation effect.
Jeddi Y, Segovia-Martin J, Servan-Schreiber E
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 41987728 · Publisher ↗
Over the past quarter century, crowdsourced forecasting has largely outperformed individual forecasters. Today, large language models (LLMs), aggregating human knowledge at scale, constitute a new form of collective inte...
Over the past quarter century, crowdsourced forecasting has largely outperformed individual forecasters. Today, large language models (LLMs), aggregating human knowledge at scale, constitute a new form of collective intelligence (CI). A central question is how LLM predictive accuracy associates with human-AI correlation, and whether this relationship exceeds what would be expected if both merely track the same underlying truth. We investigate this through the accuracy-correlation effect (ACE), which posits that as algorithmic systems improve, they increasingly correlate with human predictions, potentially diminishing human value in hybrid ensembles. Using 76 model × prompt forecast sets from 16 LLMs on 580 resolved ForecastBench questions, we computed LLM accuracy and correlations with two human aggregates (superforecasters, general public), separately for databases (n = 526) and prediction markets (n = 54) questions. Linear mixed-effects models show a robust positive association between LLM accuracy and human-AI correlation that substantially exceeds independent-errors predictions. Correlations were lower for superforecasters than for the general public, and weaker for markets than for data questions. These results support ACE while indicating that increasing correlation reflects more than improved signal tracking alone, suggesting that simultaneous increases in accuracy and correlation may reduce optimal human weights in data-rich settings, while human judgement retains critical value in contextual reasoning scenarios. This article is part of the theme issue 'The evolution of collective intelligence'.
Locals know more than it seems: a new method for revealing collective understanding, tested in three African communities.
El Aissati T, Goubraim G, Erut A … +4 more , Servan-Schreiber E, Tetteh G, Legare C, O'Madagain C
Philos Trans R Soc Lond B Biol Sci · 2026 Apr · PMID 41987727 · Publisher ↗
In the study of human cultural evolution, many theorists hold that technologies are transmitted and improved more through imitation rather than causal understanding. This view stems from results of studies that ask indiv...
In the study of human cultural evolution, many theorists hold that technologies are transmitted and improved more through imitation rather than causal understanding. This view stems from results of studies that ask individuals in isolation about technologies they use, appearing to reveal a lack of causal understanding. Here, we introduce a new method to assess the knowledge of a group that allows individuals to view their neighbours' answers before deciding what the best answer might be from among those provided by the group. We asked individuals in three farming communities from Morocco (n = 203), Mali (n = 198) and Ghana (n = 120) to explain the causal processes behind local technologies. Our method reveals that when participants can review the answers provided by their peers, the most popular final answer is more reliable than when individuals provide answers in isolation. This indicates a division of labour in how causal knowledge is stored in the community-while most individuals may have poor causal knowledge, they recognize and defer to the best answer in the group. This shows that collectively the community is more knowledgeable than methods used up to now have indicated. This article is part of the theme issue 'The evolution of collective intelligence'.
About
- Frequency
- Sun
- Papers found
- 200
- RSS feed
- Subscribe