Monday, August 04, 2025

From faces to feelings: How children learn to read emotions

Peking University

Peking University, August 4, 2025: Why do young children often miss the emotions behind adult expressions? A pioneering study led by researcher Xie Wanze from Peking University’s School of Psychological and Cognitive Sciences, in collaboration with professor Seth Pollak from the University of Wisconsin, reveals that the answer lies in a cognitive shift. Published in Nature Communications, their research shows how children aged 5-10 transition from merely “seeing” facial expressions to deeply understanding emotions, relying less on instinct and more on learned insight.

Background: The Importance of Emotion Understanding

Interpreting emotions is crucial for social bonds, yet children often struggle to decode adult feelings. This process involves perceiving facial features and applying conceptual knowledge to grasp emotional meaning. The study investigates how these cognitive mechanisms develop during childhood, filling a gap in understanding the developmental trajectory of emotion recognition.

Why it matters

As children grow, their ability to navigate complex social environments depends on a refined understanding of emotions. This research offers insights into how cognitive processes develop, with potential implications for education, parenting, and interventions for children facing social-emotional challenges.

Key findings

The study explored how children process emotions through three interconnected experiments, spanning neural activity, conceptual understanding, and behavior. In the first experiment on perception, researchers used EEG frequency tagging to show that even five-year-old children could automatically differentiate between four core facial expressions — happiness, anger, fear, and sadness — through neural responses localized in the temporo-occipital region. This perceptual ability appeared stable across different age groups. The second experiment examined conceptual knowledge through a word-similarity task, revealing that older children had more nuanced emotional associations, such as linking the word “crying” to multiple emotions, an indication of developing emotional complexity. Lastly, in the behavioral study, children participated in sorting and matching tasks. Younger participants tended to categorize expressions in broad terms of positive versus negative. At the same time, older children displayed a more refined understanding by distinguishing between specific negative emotions like anger and fear.

Core Insight: A Cognitive Shift

To integrate these findings, the team used Representational Similarity Analysis (RSA) alongside Generalized Estimating Equations (GEE) to trace the cognitive dynamics underlying emotion understanding. The results revealed a distinct developmental shift: younger children rely more heavily on perceptual cues, while older children increasingly depend on conceptual knowledge. This progression from “seeing faces” to “understanding feelings” underscores how emotional development is shaped by experience, learning, and growing cognitive sophistication throughout childhood.

Future Implications

This research highlights the dynamic interplay between perception and conceptual knowledge in children’s emotional development, offering a foundation for designing age-appropriate educational and therapeutic strategies to enhance social-emotional skills.

*This article is featured in PKU News "Why It Matters" series. More from this series.

Written by: Akaash Babar

Edited by: Zhang Jiang

Source: PKU News (Chinese)

Journal

Nature Communications

Lab to industry: InSe wafer-scale breakthrough for future electronics

Peking University

Peking University, July 30, 2025: In a landmark advancement for next-generation electronics, researchers from the International Center for Quantum Materials at Peking University in collaboration with Renmin University of China have successfully fabricated wafer-scale two-dimensional indium selenide (InSe) semiconductors. Led by Professor Liu Kaihui, the team developed a novel “solid–liquid–solid” growth strategy that overcomes long-standing barriers in 2D semiconductor manufacturing.

Published in Science under the title “Two-dimensional indium selenide wafers for integrated electronics,” the study demonstrates exceptional electronic performance, surpassing all previously reported 2D film-based devices. The fabricated InSe transistors exhibit ultra-high electron mobility and a near-Boltzmann-limit subthreshold swing at room temperature, establishing a new benchmark for 2D semiconductors.

Background: Why InSe?
Indium selenide, often referred to as a “golden semiconductor,” offers an ideal combination of properties—low effective mass, high thermal velocity, and a suitable bandgap. Despite these advantages, its wafer-scale integration has remained elusive due to the difficulty of precisely maintaining a 1:1 atomic ratio between indium and selenium during synthesis. Traditional methods have only yielded microscopic flakes, insufficient for practical electronic applications.

Why it matters
As Moore’s Law slows and silicon nears its physical limits, the semiconductor industry faces growing pressure to identify alternative channel materials. In this context, the successful fabrication of large-area crystalline InSe wafers represents a pivotal step toward faster, more energy-efficient, and smaller chips for next-generation electronics.

The In–Se system faces challenges due to multiple stable phases and extreme vapor pressure differences between indium and selenium, making it difficult to maintain stoichiometry during growth. These issues hinder phase purity, crystal quality, and overall device stability. Professor Liu Kaihui’s team developed a novel solid–liquid–solid conversion strategy. This process begins with the deposition of an amorphous InSe thin film onto sapphire substrates using magnetron sputtering. The wafer is then encapsulated with low-melting-point indium and sealed inside a quartz cavity. When heated to approximately 550 °C, the indium creates a localized, indium-rich environment that promotes controlled dissolution and recrystallization at the interface. This carefully orchestrated reaction results in the formation of uniform, single-phase crystalline InSe films. This method produced 2-inch wafers with world-first crystallinity, phase purity, and thickness uniformity for 2D InSe.

Device Performance
Using these wafers, the team fabricated large-scale transistor arrays that demonstrated outstanding performance, including an electron mobility of up to 287 cm²/V·s and an average subthreshold swing of 67 mV/dec. The devices exhibited excellent behavior at sub-10 nm gate lengths, characterised by reduced drain-induced barrier lowering (DIBL), lower operating voltages, enhanced on/off current ratios, and efficient ballistic transport at room temperature.

Significantly, the devices surpassed 2037 IRDS projections for delay and energy-delay product (EDP), positioning InSe ahead of silicon in key future benchmarks.

This breakthrough opens a new pathway for the development of next-generation, high-performance, low-power chips, which are expected to be applied widely in cutting-edge fields such as artificial intelligence, autonomous driving, and smart terminals in the future. Reviewers of Science have hailed this work as "an advancement in crystal growth."

*This article is featured in PKU News "Why It Matters" series. More from this series.
Read more: https://www.science.org/doi/10.1126/science.adu3803

Written by: Akaash Babar
Edited by: Zhang Jiang
Source: Xinhua News

Journal

Science

Crop monitoring system utilizing IoT, AI and other tech showcased at ASABE

South Dakota State University

Smart Crop Monitoring System — image:
Researchers from South Dakota State University developed a high-tech system to help farmers optimize crop yields while lowering costs. The system provides secure monitoring and data collection powered by solar energy, as shown here.
view more
Credit: Lin Wei, South Dakota State University

Researchers from South Dakota State University presented a high-tech system to help farmers optimize crop yields while lowering costs at the 2025 annual meeting of the American Society of Agricultural and Biological Engineers. The system, detailed in Integrating IoT and secure data transmission in a crop monitoring system, tracks and analyzes crop development through data collected by sensors, biosensors, the Internet of Things and AI.

While the majority of projects that build systems utilizing IoT only simulate post-quantum security on super computers, the work presented by SDSU Professor Lin Wei and his Ph.D. student Manish Shrestha implemented security in a real-world end-to-end, sensor-to-cloud, application.

“This work demonstrated that strong, future-proof security can run directly on small devices, potentially eliminating the need for large servers to protect IoT data,” Shrestha said. “This ensures farming data remains private, verifiable and resilient—even against future quantum computer attacks.”

Protecting farmer’s data collected by the system was imperative to Shrestha and the team, who used advanced security protocols, encryption and cryptography to ensure the massive amount of data was safe while being stored and analyzed in the cloud. Included in the data were measurements of soil conditions such as temperature, humidity and available nutrients; potential plant stresses such as nutrient deficiencies, disease presence and pest threats; and environmental factors.

Once data from all of these measurements was analyzed by the researchers, it was presented to farmers, allowing them to make more informed decisions on their management practices including irrigation, fertilization, disease and pest control, without risking their operation’s information.

The importance of heightened cybersecurity practices in agriculture applications was a hot topic at the meeting.

“Our research received considerable attention, with many experts emphasizing how cybersecurity must be a core component when developing smart farming technologies,” Shrestha, said. “There was a common thread of people recognizing the need for a secure infrastructure for all the data farmers are collecting.”

The research team plans to improve their system in the near future by speeding up sensor data processing and using a solar powered battery rather than a chemical one to lengthen the time between charges among other updates.

Journal

Transactions of the ASABE

DOI

10.13031/aim.202500564

Method of Research

Observational study

Article Title

Integrating IoT and secure data transmission in a crop monitoring system

Pan Feng’s team advances inverse design of catalytic materials with topological AI

Peking University

Peking University, June 18, 2025: A collaborative research team led by Professor Pan Feng from the School of New Materials at Peking University Shenzhen Graduate School has developed a topology-based variational autoencoder framework (PGH-VAEs) to enable the interpretable inverse design of catalytic active sites. Their study, titled “Inverse design of catalytic active sites via interpretable topology-based deep generative models” and published in npj Computational Materials, introduces a novel integration of graph-theoretic structural chemistry, algebraic topology, and deep generative models, enabling the rational design of catalysts with targeted adsorption properties from performance objectives.

Background
Designing catalysts with precise active sites is crucial for enhancing efficiency in energy and chemical processes. Traditional forward-design methods, based on DFT and machine learning, struggle with complex systems, such as high-entropy alloys (HEAs), due to their limited interpretability and data constraints. Graph-based representations of atomic structures, paired with persistent GLMY homology (PGH), a topological tool for asymmetric graphs, provide a new approach to analyzing and generating catalytic structures.

Why it matters
This interpretable inverse design framework provides a powerful alternative to trial-and-error methods in catalyst discovery. This work demonstrates that interpretable inverse design is no longer out of reach. By linking topological descriptors with physical performance metrics, the framework provides a transparent pathway from theoretical modeling to practical catalyst synthesis. Such breakthroughs are especially crucial for HEAs and other structurally complex catalysts where trial-and-error experimentation is costly and inefficient.

Key Findings
In this groundbreaking study, the researchers introduced a physically interpretable inverse design framework that combines graph-theoretic structural representations with topological analysis and deep generative modeling. Using persistent GLMY homology (PGH), they extracted topological invariants, such as atomic connectivity and structural voids, from complex catalytic configurations, enabling a deeper understanding of how local and long-range structural features influence catalytic performance. To capture these interactions, they designed a dual-channel representation system that separately encodes atomic coordination and distant elemental modulation effects. This data was then used to train a variational autoencoder (VAE) coupled with a gradient boosting regressor (GBRT), achieving highly accurate predictions of *OH adsorption energy, with a mean absolute error of just 0.045 eV, despite being trained on a relatively small dataset of around 1100 DFT samples.

Remarkably, the study uncovered a strong linear correlation between topological descriptors, particularly Betti numbers, and adsorption properties, providing rare physical insight into the structure–performance relationship. The model also successfully generated optimal active site structures in IrPdPtRhRu high-entropy alloys, identifying Pt/Pd as preferred bridge atoms and Ru as a distant regulator. Furthermore, it predicted ideal compositional ratios for different crystal surfaces, offering precise and actionable targets for experimental validation.

Future Implications
This study sets a new benchmark for interpretable, data-driven materials design. Though focused on HEAs, the framework can be extended to other catalysts and materials for energy, environmental, and industrial applications, offering a scalable path to rational, AI-guided material discovery.

*This article is featured in PKU News "Why It Matters" series. More from this series.
Read more: https://www.nature.com/articles/s41524-025-01649-8

Written by: Akaash Babar
Edited by: Zhang Jiang
Source: PKU News (Chinese)

Journal

npj Computational Materials

A comparison of colorful hamlets from the Caribbean challenges ideas about how species arise

Smithsonian Tropical Research Institute

Toddlers can name a few animals. Older kids group animals into categories (birds, fish). And teenagers can sketch a rough tree of life. But when 16 grown-up biologists—five of them affiliated with the Smithsonian Tropical Research Institute—try to explain why colorful reef fishes called hamlets are different species, it gets complicated. Their results in the journal Science Advances challenge common explanations about using genetic differences to tell species apart.

“Most studies explaining how the different species in a group evolve begin by showing a family tree based on the genetic differences between the species,” said Oscar Puebla, senior author, STRI Research Associate, and professor at the Leibnitz Center for Tropical Marine Research (ZMT) in Germany. “But in this case, there is just one genetic split among all 19 species of hamlets.”

Species form when animals adapt to new conditions. Animals that can’t adapt die out and the ones that can adapt reproduce, becoming so genetically different that they can no longer interbreed with other animals in the original group. Maybe the animals were isolated in space, like on the opposite sides of a mountain range; or their paths never crossed because they were active at different times of day…but overall, no matter the cause, they became genetically different. In the early stages of species differentiation, animals may still interbreed and only differ at genes that are directly relevant to adaptation—the genic view of speciation. In principle, these genes should allow researchers to reconstruct a family tree.

“To our surprise, when we looked at the genetic data, we realized that no single gene allows us to reconstruct a family tree for the group,” said Martin Helmkampf, senior researcher at ZMT. “And we are sure of this because we screened the entire genomes of 335 fishes. This challenges how people think about what species are and how they arise.”

Co-authors Floriane Coulmance, the first winner of the D. Ross Robertson Postdoctoral Fellowship for Field Studies on Neotropical Reef Fishes, now at the University of Lausanne in Switzerland, and Melanie Heckwolf, former post-doctoral fellow at STRI, developed an innovative underwater camera with a color chart, making it possible to describe the colors of living fish pixel by pixel and to compare colors even when the fish were photographed under very different light conditions. Read about that here in our web story, “Living Color.”

Later, researchers in the group sequenced the entire genomes of 335 fish. Based on the color data and the genetic data, they identified just one gene that seems to be involved in species differences. This gene, called casz1 is expressed in the skin, eyes and brain of hamlets and it is probably responsible for determining color patterns and mate choice. However, even this gene does not allow researchers to reconstruct a family tree for the group. This is probably because species differences are encoded by many genes that act in concert.

“So, in the end, we have to live with the fact that in some cases it is impossible to reconstruct a family tree that differentiates the species,” said Helmkampf.

There is a very similar explanation for speciation in the Heliconius butterflies, studied at the Smithsonian’s laboratories in Gamboa, Panama. As Owen McMillan, STRI staff scientist and coauthor of the study says: “If there is one thing that we’ve learned from our ability to sequence the genomes of many individuals, it’s just how fluid the boundaries are between what we call species. Hybridization among species occurs a lot and it is turning out to be an important way to rapidly evolve new traits and/or exploit new environments in everything from humans to butterflies, birds, and coral reef fishes. It is also a way to rapidly evolve new species, something we see in butterflies, and is probably occurring in hamlets. Of course this makes sorting things into defined groups challenging, but it is a remarkable testament to how evolution works to create Earth’s biodiversity.”

Affiliations of the 16 authors include: The Leibnitz Center for Tropical Marine Research (ZMT); the Institute for Chemistry and Biology of the Marine Environment (ICBM); the Carlo von Ossietsky Universität Oldenburg; the Smithsonian Tropical Research Institute (STRI); the Instituto para el Estudio de las Ciencias del Mar (CECIMAR); the Universidad Nacional de Colombia sede Caribe; the Corporation Center of Excellence in Marine Science (CEMarin); the Division of Biosciences, Faculty of Life Science, University College London (UCL); the Senckenberg Research Institute and Natural History Museum, Frankfurt am Main; the LOEWE Centre for Translational Biodiversity Genomics, Frankfurt am Main, Germany; the Wellcome Sanger Institute, Tree of Life, Wellcome Genome Campus, Cambridge; the Laboratorio de Biología Acuática, Facultad de Biología, Universidad Michoacana de San Nicolás Hidalgo; el Laboratorio Nacional de Análisis y Síntesis Ecológica para la Conservación de Recursos Genéticos de México, Escuela Nacional de Estudios Superiores, Unidad Morelia, Universidad Nacional Autónoma de México, Morélia; the Department of Plant and Wildlife Sciences, Brigham Young University; the Institute of Clinical Molecular Biology (IKMB), Kiel University; School of Biological Sciences, The University of Oklahoma, Norman; el Departamento de Biología, Universidad del Valle, Cali, Colombia; the Ocean Science Foundation, Irvine; the Guy Harvey Research Institute; the Scripps Institution of Oceanography, University of California.

About the Smithsonian Tropical Research Institute
Headquartered in Panama City, Panama, STRI is a unit of the Smithsonian Institution in Washington, DC, USA. Our mission is to understand tropical biodiversity and its importance to human welfare, to train students to conduct research in the tropics and to promote conservation by increasing public awareness of the beauty and importance of tropical ecosystems. Watch our video, and visit our website, Facebook, X and Instagram for updates.

Reference: Helmkampf M, Coulmance F, Heckwolf MJ, Acero A, Balard A, Bista I, Dominguez O, Frandsen PB, Torres-Oliva M, Santaquieteria A, Tavera J, Victor BC, Robertson DR, Betancur-R R, McMillan WO, Puebla O. 2025. Radiation with reproductive isolation in the near absence of phylogenetic signal. Science Advances 11(30), eadt0973. https://www.science.org/doi/full/10.1126/sciadv.adt0973

Journal

Science Advances

DOI

10.1126/sciadv.adt0973

Article Title

Radiation with reproductive isolation in the near-absence of phylogenetic signal

Tracing brain chemistry across humanity’s family tree

Small changes to an enzyme suggest how modern humans differ from Neanderthals and Denisovans in biochemistry and behavior

Okinawa Institute of Science and Technology (OIST) Graduate University

ADSL enzymatic activities and protein sequences — image:
In the study, the researchers examined the activity of ADSL by measuring the concentration of protein substrates SAICAr and S-Ado (in red), which increase when ADSL activity decreases. Below, the amino acid sequence differences among modern humans, Neanderthals and Denisovans, chimpanzees, and mice are shown. The change at position 429 is the only difference between modern humans and ancestral hominins. The substitution in mice on position 428 (in red) was not found to alter the activity of ADSL.
view more
Credit: Ju et al., 2025

The evolutionary success of our species may have hinged on minute changes to our brain biochemistry after we diverged from the lineage leading to Neanderthals and Denisovans about half a million years ago

Two of these tiny changes that set modern humans apart from Neanderthals and Denisovans affect the stability and genetic expression of the enzyme adenylosuccinate lyase, or ADSL. This enzyme is involved in the biosynthesis of purine, one of the fundamental building blocks of DNA, RNA, and other important biomolecules. In a study now published in PNAS, researchers from the Okinawa Institute of Science and Technology (OIST), Japan and the Max Plank Institute for Evolutionary Anthropology, Germany have discovered that these changes may play an important role in our behavior, contributing new pieces to the great puzzle of who we humans are and where we come from. “Through our study, we have gotten clues into the functional consequences of some of the molecular changes that set modern humans apart from our ancestors,” says first author Dr. Xiang-Chun Ju of the Human Evolutionary Genomics Unit at OIST.

The ADSL enzyme is made up of a chain of 484 amino acids. The modern and ancestral variants of this enzyme differ by just one of these amino acids: at position 429, the alanine in the ancestral form has been substituted with a valine in the modern. In vitro, this change has been observed to reduce the stability of the protein. The team has now shown that in mouse models, this translates to higher concentrations of the substrates that ADSL catalyzes in several organs, especially in the brain.

Given that genetic ADSL deficiency is known to cause psychomotor retardation and cognitive impairments in humans, the researchers explored the possible behavioral effects of this substitution. In an experimental setup where water is made available to mice following a visual or sound cue, they found that female mice with the substitution consistently accessed water more frequently than their littermates when they were thirsty, suggesting that the reduced activity of the enzyme allowed them to better compete for a scarce resource.

The amino acid substitution is absent in both Neanderthals and Denisovans but present in virtually all present-day humans, showing that this change must have appeared in modern humans after they separated from the lineage leading to Neanderthals and Denisovans but before they left Africa. “It’s too early to translate these findings directly to humans, as the neural circuits of mice are vastly different,” adds Dr. Ju. “But the substitution might have given us some evolutionary advantage in particular tasks relative to ancestral humans.”

The team next searched for other related genetic changes that might affect ADSL activity in present-day humans. They identified a set of genetic variants in a non-coding region of the ADSL gene, which are present in at least 97% of all present-day human genomes. Statistical tests involving Neanderthal, Denisovan, and modern African, European, and East Asian genetic sequences provided strong evidence that these variants have been positively selected among modern humans.

Fascinatingly, the researchers found that rather than compensating for the reduced activity of ADSL caused by the amino acid change, the non-coding changes reduce ADSL RNA expression, further diminishing its activity – again, especially in the brain. “This enzyme underwent two separate rounds of selection that reduced its activity – first through a change to the protein’s stability and second by lowering its expression. Evidently, there’s an evolutionary pressure to lower the activity of the enzyme enough to provide the effects that we saw in mice, while keeping it active enough to avoid ADSL deficiency disorder,” explains co-author Dr. Shin-Yu Lee of the same research unit at OIST.

“Our results open up many questions,” explains Professor Izumi Fukunaga of the Sensory and Behavioral Neuroscience Unit at OIST. “For example, it’s unclear why only female mice seemed to gain a competitive advantage. Behavior is complex. Accessing water proficiently involves processing sensory information, learning which actions lead to rewards, navigating social interactions, motor planning, and many other processes. Each of these may involve multiple brain regions. As such, more studies are needed to understand the role of ADSL in behavior.”

Professor Svante Pääbo, leader of the Human Evolutionary Genomics Unit, summarizes: “There are a small number of enzymes that were affected by evolutionary changes in the ancestors of modern humans. ADSL is one of them. We are beginning to understand the effects of some of these changes, and thus to puzzle together how our metabolism has changed over the past half million years of our evolution. A next step will be to study what effects combinations of these changes may have.”

The global distribution of the set of gene variants that reduce ADSL expression, with percentages indicating the frequency of the modern human variants in red. As each person carries two copies of the ADSL gene, about 97% of everybody today carries at least one copy of the modern variants.

Credit

Ju et al., 2025

Journal

Proceedings of the National Academy of Sciences

DOI

10.1073/pnas.2508540122

Method of Research

Experimental study

Subject of Research

Not applicable

Article Title

The activity and expression of adenylosuccinate lyase were reduced during modern human evolution, affecting brain and behavior

Article Publication Date

4-Aug-2025

LA REVUE GAUCHE - Left Comment

Monday, August 04, 2025

From faces to feelings: How children learn to read emotions

Journal

Lab to industry: InSe wafer-scale breakthrough for future electronics

Journal

Crop monitoring system utilizing IoT, AI and other tech showcased at ASABE

Journal

DOI

Method of Research

Article Title

Pan Feng’s team advances inverse design of catalytic materials with topological AI

Journal

A comparison of colorful hamlets from the Caribbean challenges ideas about how species arise

Journal

DOI

Article Title

Tracing brain chemistry across humanity’s family tree

The global distribution of the set of gene variants that reduce ADSL expression, with percentages indicating the frequency of the modern human variants in red. As each person carries two copies of the ADSL gene, about 97% of everybody today carries at least one copy of the modern variants.

Credit

Journal

DOI

Method of Research

Subject of Research

Article Title

Article Publication Date

About Me