can surface in training data even in objects as “universal” as chairs, as observed in these doodle patterns on the left. The chart on the right shows how we uncovered in standard open source data sets such as ImageNet. Undetected or uncorrected, such biases may strongly influence model behavior.
What is brain health and why is it important, yongjun wang.
1 Department of Neurology, Beijing Tiantan Hospital, Capital Medical University, Beijing, China
2 China National Clinical Research Center for Neurological Diseases, Beijing, China
Yongjun Wang and colleagues discuss the definition of brain health and the opportunities and challenges of future research
The human brain is the command centre for the nervous system and enables thoughts, memory, movement, and emotions by a complex function that is the highest product of biological evolution. Maintaining a healthy brain during one’s life is the uppermost goal in pursuing health and longevity. As the population ages, the burden of neurological disorders and challenges for the preservation of brain health increase. It is therefore vital to understand what brain health is and why it is important. This article is the first in a series that aims to define brain health, analyse the effect of major neurological disorders on brain health, and discuss how these disorders might be treated and prevented.
Currently, there is no universally recognised definition of brain health. Most existing definitions have only a general description of normal brain function or emphasise one or two dimensions of brain health. The US Centers for Disease Control and Prevention defined brain health as an ability to perform all the mental processes of cognition, including the ability to learn and judge, use language, and remember. 1 The American Heart Association/American Stroke Association (AHA/ASA) presidential advisory defined optimal brain health as “average performance levels among all people at that age who are free of known brain or other organ system diseases in terms of decline from function levels, or as adequacy to perform all activities that the individual wishes to undertake.” 2
The brain is a complex organ and has at least three levels of functions that affect all aspects of our daily lives: interpretation of senses and control of movement; maintenance of cognitive, mental, and emotional processes; and maintenance of normal behaviour and social cognition. Brain health may therefore be defined as the preservation of optimal brain integrity and mental and cognitive function at a given age in the absence of overt brain diseases that affect normal brain function.
Several neurological disorders may disrupt brain function and affect humans’ health. Medically, neurological disorders that cause brain dysfunction can be classified into three groups:
These neurological disorders may have different or common effects on brain health and function. For instance, Alzheimer’s disease is the main type of dementia, with a decline in different domains of cognitive function. Mood disorders may cause dysfunction in execution, reward processing, and emotional regulations. In addition to physical disability, aphasia, gait and balance problems, and cerebrovascular diseases may lead to cognitive impairment and dementia, which are neglected by both patients and physicians.
Human ageing is mainly reflected in the aspects of brain ageing and degradation of brain function. The number of people aged 60 years and over worldwide was around 900 million in 2015 and is expected to grow to two billion by 2050. 3 With the increases in population ageing and growth, the burden of neurological disorders and challenges to the preservation of brain health steeply increase. People with neurological disorders will have physical disability, cognitive or mental disorders, and social dysfunction and be a large economic burden.
Globally, neurological disorders were the leading cause of disability adjusted life years (276 million) and the second leading cause of death (9 million) in 2016, according to the Global Burden of Diseases study. 4 Stroke, migraine, Alzheimer’s disease and other dementias, and meningitis are the largest contributors to neurological disability adjusted life years. 4 About one in four adults will have a stroke in their lifetime, from the age of 25 years onwards. 5 Roughly 50 million people worldwide were living with dementia in 2018, and the number will more than triple to 152 million by 2050. 6 In the following decades, governments will face increasing demand for treatment, rehabilitation, and support services for neurological disorders.
Opportunities and challenges exist in the assessment of brain health, the mechanism of brain function and dysfunction, and approaches to promote brain health ( box 1 ).
Defining and promoting optimal brain health require the scientific evaluation of brain health. However, it is difficult to comprehensively evaluate or quantify brain health through one metric owing to the multidimensional aspects of brain health. Many structured or semistructured questionnaires have been developed to test brain health by self-assessments or close family member assessments of daily function or abilities. In recent decades new structural and functional neuroimaging techniques have been applied to evaluate brain network integrity and functional connectivity. 7 However, these subjective or objective measures have both strengths and weaknesses. For instance, scales such as the mini-mental state examination and Montreal cognitive assessment are simple and easy to implement but are used only as global screening tools for cognitive impairment; tests such as the digit span, Rey-Osterrieth complex figure test, trail making A and B, Stroop task, verbal fluency test, Boston naming test, and clock drawing test are used mainly to assess one or two specific domains of memory, language, visuospatial, attention, and executive function; and neuroimaging techniques, although non-invasive and objective, still have disadvantages of test contraindications, insufficient temporal or spatial resolution, motion artefact, and high false discovery rates, which limit their clinical transformation.
Another difficulty in measuring brain health is that age, culture, ethnicity, and geography specific variations exist in the perception of optimal brain health. Patient centred assessment of brain function, such as self-perception of cognitive function and quality of life, should also be considered when measuring brain health. 8 Universal acceptable, age appropriate, multidimensional, multidisciplinary, and sensitive metrics or tools are required to comprehensively measure and monitor brain function and brain health.
To promote optimal brain health, we need a better understanding of the mechanisms of brain function and dysfunction. Unfortunately, little is known about the working mechanism of the brain. Although we have made considerable developments in neuroscience in recent decades, we still cannot totally decipher the relations between spatiotemporal patterns of activity across the interconnected networks of neurons and thoughts or the cognitive and mental state of a person. 9 Recent progress in brain simulation and artificial intelligence provides a vital tool to understand biological brains, and vice versa. 10 11 The development of brain inspired computation, brain simulation, and intelligent machines was emphasised in the European Union and China Brain Project. 9 12
Meanwhile, the mechanisms behind the brain dysfunction in some neurological disorders are still not well understood, especially for mental and neurodegenerative disorders. Further investigation of the mechanisms of brain diseases may indicate approaches to treatment and improve brain function. Brain imaging based cognitive neuroscience may unravel the underlying brain mechanism of cognitive dysfunction and provide an avenue to develop a biological framework for precision biomarkers of mood disorders. 13
Most common neurological diseases, such as cerebrovascular diseases and Alzheimer’s disease, have complex aetiopathologies, typically involving spatial-temporal interactions of genetic and environmental factors. However, a single genetic factor could account for the disease progression of monogenic neurological disorders. These diseases could be more readily investigated by simplified cross species modelling, leading to better understanding of their mechanisms and greater efficiency in testing innovative therapies. Such research may provide a window to promote the investigation of common neurological disorders and general brain health, as discussed by Chen and colleagues elsewhere in this series. 14
Few effective approaches are available to prevent and treat brain dysfunction in some major neurological disorders, such as dementia. Neurons are not renewable, and brain dysfunction is always irreversible. Recent trials targeting amyloid clearance and the selective inhibition of tau protein aggregation failed to improve cognition or modify disease progression in patients with mild Alzheimer’s disease. 15 16 More attention has focused on other potential therapeutic targets, such as vascular dysfunction, inflammation, and the gut microbiome, as discussed by Shi and colleagues. 17 In particular, recent studies showed that the early impairment of cognition was induced by the disruption of neurovascular unit integrity, which may cause hypoperfusion and the breakdown of the blood-brain barrier and subsequent impairment in the clearance of proteins in the brain. 18 19 Physical activity, mental exercise, a healthy diet and nutrition, social interaction, ample sleep and relaxation, and control of vascular risk factors are considered six pillars of brain health. The AHA/ASA presidential advisory recommended the AHA’s Life’s Simple 7 (non-smoking, physical activity, healthy diet, appropriate body mass index, blood pressure, total cholesterol, and blood glucose) to maintain optimal brain health. 2 Pan and colleagues discuss how this may indicate a new dawn of preventing some cognitive impairment and brain dysfunction by preventing vascular risk factors or cerebrovascular diseases. 20
For other neurological disorders with potential therapeutic approaches, the main aim is to preserve brain function. Impaired brain function due to anatomical structural damage is underestimated in patients with neurosurgical diseases such as brain tumours, trauma, and epilepsy. In recent years, treatment targets for neurosurgical diseases have changed from focusing on survival or life expectancy to balancing brain structures and functions. Precise preservation of brain function requires an understanding of the exquisite relation between brain structure and function and advanced technologies to visualise brain structure-function relations. 21
Another example of the predicament associated with protection of brain function is uncertainty in the treatment response in epilepsy management. Current standard care for epilepsy relies on a trial and error approach of sequential regimens of antiseizure medications. The time delay due to this treatment approach means that such treatments may be less effective and irreversible damage may occur. Chen and colleagues 22 describe how recent advances in personalised epilepsy management based on artificial intelligence, genomics, and patient derived stem cells are bringing some hope to overcome this predicament in epilepsy management and promise a more effective strategy. 23 24
Brain health is the maintenance of multidimensional aspects of brain function. However, several neurological disorders may affect brain health in one or more aspects of brain function. Deciphering and promoting the function and health of the brain, the most mysterious organ in the human body, will have a dramatic impact on science, medicine, and society. 25 In the past seven years, a number of large scale brain health initiatives have been launched in several countries to promote the development of neuroscience, brain simulation, and brain protection. 9 However, further challenges are raised by the different key research directions of brain projects in different countries. In the face of these challenges, Liu and colleagues argue that collaboration on brain health research is urgently needed. 26 As the other articles in this series describe, coordinated research has enormous potential to improve the prognosis of brain disorders.
Contributors and sources: YW proposed the idea for this series on brain health. YW and YP drafted the first manuscript. All the authors critically reviewed and revised the manuscript. YP and HL expertise is in the area of clinical research methods and clinical research on stroke. YW is an expert in clinical research on stroke and neurological diseases. YW is the guarantor.
Competing interests We have read and understood BMJ policy on declaration of interests and declare that the study was supported by grants from the National Science and Technology Major Project (2017ZX09304018), National Key R&D Program of China (2018YFC1312903, 2017YFC1310902, 2018YFC1311700, and 2018YFC1311706), National Natural Science Foundation of China (81971091), Beijing Hospitals Authority Youth Programme (QML20190501), and Beijing Municipal Science and Technology Commission (D171100003017002).
Provenance and peer review: Commissioned; externally peer reviewed.
This article is part of a series launched at the Chinese Stroke Association annual conference on 10 October 2020, Beijing, China. Open access fees were funded by the National Science and Technology Major Project. The BMJ peer reviewed, edited, and made the decision to publish these articles.
Latest news.
Discover our latest AI breakthroughs and updates from the lab
View all posts
FermiNet: Quantum physics and chemistry from first principles
Using deep learning to solve fundamental problems in computational quantum chemistry and explore how matter interacts with light
A new generation of African talent brings cutting-edge AI to scientific challenges
Food security, healthcare and exploring the cosmos are among the ways students of a new pan-African Master’s program aspire to apply AI
Responsibility & Safety
Mapping the misuse of generative AI
New research analyzes the misuse of multimodal generative AI today, in order to help build safer and more responsible technologies.
New episodes out now
Breakthrough research. Transformative products.
View all technologies
A universal AI agent that is helpful in everyday life
The most general and capable AI models we've ever built.
Our most capable generative video model.
Our highest quality text-to-image model.
Robust and scalable tool for watermarking and identifying AI-generated images.
Breakthrough AI system accurately predicts the 3D models of protein structures — and accelerates research in nearly every field of biology.
We want AI to benefit the world, so we must be thoughtful about how it’s built and used.
We help anticipate a broad spectrum of AI-related risks, explore ways of preventing them from happening, and find ways to address them if they do.
AI Principles
While we are optimistic about the potential of AI, we recognize that advanced technologies can raise important challenges that must be addressed clearly, thoughtfully, and affirmatively.
Gateway to research with smart searches, audio podcasts from papers, easy uploads, and interactive chats.
What is paperbrain, is it free, where do we source our papers from.
Pni innovator awards spur new research on ai, the brain, and hormones.
For the first time, two senior postdoctoral researchers earned the annual Innovator Awards this year at the Princeton Neuroscience Institute (PNI), along with the traditional two awards reserved for pairs of collaborating interdisciplinary faculty members.
“There’s a dearth of funding opportunities for senior postdoctoral researchers,” said Mala Murthy, Ph.D. , PNI's director and the Karol and Marnie Marcin ’96 professor of neuroscience. “By opening up PNI’s Innovator Awards to postdoctoral scholars, we aim to support exciting new research directions undertaken in collaboration with their advisors, and to facilitate transitions to the next career stage.”
PNI postdoctoral research associate Rich Pang, Ph.D., and associate research scholar Chris Langdon, Ph.D. are the inaugural postdoc innovator awardees, as well as faculty member duos Uri Hasson, Ph.D. and Casey Lew-Williams, Ph.D., and Andrew Leifer, Ph.D. and Joshua Shaevitz, Ph.D.
Pang works across the labs of Dr. Murthy, PNI professor Jonathan Pillow, Ph.D. , and associate PNI faculty member and the John Archibald Wheeler/Battelle professor of physics William Bialek, Ph.D. Pang’s project will use recurrent neural networks to analyze troves of whole-brain activity recordings to better understand how neural activity propagates throughout the brain and how it produces behavior.
“All of the neural data I have studied to date has been from recordings of small numbers of neurons or in very limited conditions,” Pang wrote in his application. “This proposal is a new direction into the analysis and modeling of large-scale recordings across a variety of complex conditions.”
Inspired by recent work from the lab of Hakan Türeci, Ph.D. , a theoretical physicist and professor of electrical and computer engineering at Princeton, Pang will explore if Türeci’s newly developed framework for charting nonlinear dynamic systems, dubbed “eigentests,” can further reveal the computational properties of neural networks.
Similar to Pang, Langdon, an associate research scholar in the lab of PNI associate professor Tatiana Engel, Ph.D. , will use the Innovator Award funds to support his work on using recurrent neural networks to bridge our understanding of brain circuit connectivity, neural activity, and behavior.
“Chris’s project has the potential to transform our understanding of how cognitive functions arise from dynamic interactions in neural circuits,” Engel wrote about Langdon’s proposal. “Confirmation of theoretical predictions about the existence of functional cell types and their relationship to the neural response dimensionality may reveal a universal organizing principle of cortical circuits.”
The old parenting adage goes that when you start a family and have kids, “the days are long, but the years are short”.
For PNI and psychology professor Uri Hasson, Ph.D. and associated PNI faculty member Casey Lew-Willaims, Ph.D. , each day is about 1,500 hours long and stretches out until a child’s third birthday.
That’s because Hasson and Lew-Williams will use their funds to support their unprecedented research that tracks children from the day they arrive home from the hospital until their third birthday.
Multiple cameras and microphones are installed across different rooms in each participating family’s house to track the development of fifteen babies in their natural home environment. In total, each three-year-old will have the most comprehensive and detailed home videos ever collected of the first 1000 days of their lives.
In doing so, Hasson and Lew-Williams aim to gain a better understanding of how children naturally learn language by interacting with their supportive social environments.
After resolving various ethical and technical obstacles, Hasson and Lew-Williams have recruited a diverse group of 15 families across New Jersey and Pennsylvania, including mixed-race, multigenerational, and LGBTQ+ families, which reflects each state’s demographics.
The Innovator Award funds will help support the team’s current recording endeavors and help build on their deep learning pipeline to sort through the petabytes of data collected across the project.
“With 100,000 one-minute audiovisual clips uploaded daily, we are collecting 36.5 million minutes annually, making the dataset too large to be manually labeled by human annotators,” the team wrote in their application. “The machine learning module will enable our team to quantify, for the first time, the natural, everyday statistics in infants’ environments that give rise to learning.”
Furthermore, their data will enable them to build a new generation of large language models capable of learning language from the child-centered linguistic input that each child receives in their natural environments.
The pioneering neuroscientist Eve Marder often quips that neural circuit diagrams, like connectomes, are “…absolutely necessary but completely insufficient for understanding nervous system function.”
Missing from the brain’s road map is a fuller understanding of the traffic patterns, speed limits, and other cues that dictate how cells talk to one another. Neuroscientists have often focused on classic chemical messengers, like GABA and glutamate, but those neurotransmitters are only one of many modes of communication.
Unlike neurotransmitters, which go from one neuron to the next in like, neuropeptides are released from large packets that can travel many cells away from its host. That makes understanding who each cell is trying to communicate with much trickier to track.
To address this, PNI and physics associate professor Andrew Leifer, Ph.D. , and associated PNI professor Joshua Shaevitz, Ph.D. will collaborate for their Innovator Award to better understand where neuropeptides and what rules govern their communication style by studying such phenomena in the roundworm Caenorhabditis elegans . C. elegans , as it’s often abbreviated, has a relatively simple nervous system totaling 302 well understood and mapped neurons.
Leifer and Shaevitz aim to take advantage of the roundworm’s manageable roadmap by adding a new layer of understanding how peptides travel along its neural circuitry.
“We propose to use sophisticated new peptide sensors and volumetric functional imaging to directly measure brain-wide neuropeptide dynamics in C. elegans, including where peptides go once released, how quickly they diffuse, and how anatomy influences their dynamics,” Leifer and Shaevitz wrote in their proposal. “The speed, extent and principles that govern how peptides travel will provide insights into their role in neural function.”
The Innovator Awards are generously supported by Endowments from the McDonnell Center for Systems Neuroscience, the Bezos Center for Neural Circuit Dynamics, and the Scully Center for the Neuroscience of Mind and Behavior.
Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
You can also search for this author in PubMed Google Scholar
You have full access to this article via your institution.
Monkeypox virus particles (shown in this coloured electron micrograph) can spread through close contact with people and animals. Credit: NIAID/Science Photo Library
When the World Health Organization (WHO) declared a public-health emergency over mpox earlier this month , it was because a concerning form of the virus that causes the disease had spread to multiple African countries where it had never been seen before. Since then, two people travelling to Africa — one from Sweden and one from Thailand — have become infected with that type of virus, called clade Ib, and brought it back to their countries.
Monkeypox virus: dangerous strain gains ability to spread through sex, new data suggest
Although researchers have known about the current outbreak since late last year, the need for answers about it is now more pressing than ever. The Democratic Republic of the Congo (DRC) has spent decades grappling with monkeypox clade I virus — the lineage to which Ib belongs. But in the past, clade I infections usually arose when a person came into contact with wild animals, and outbreaks would fizzle out.
Clade Ib seems to be different, and is spreading largely through contact between humans, including through sex . Around 18,000 suspected cases of mpox, many of them among children, and at least 600 deaths potentially attributable to the disease have been reported this year in the DRC alone.
How does this emergency compare with one declared in 2022, when mpox cases spread around the globe? How is this virus behaving compared with the version that triggered that outbreak, a type called clade II? And will Africa be able to rein this one in? Nature talks to researchers about information they are rushing to gather.
It’s hard to determine, says Jason Kindrachuk, a virologist at the University of Manitoba in Winnipeg, Canada. He says that the DRC is experiencing two outbreaks simultaneously. The clade I virus, which has been endemic in forested regions of the DRC for decades, circulates in rural regions, where people get it from animals. That clade was renamed Ia after the discovery of clade Ib. Studies in animals suggest that clade I is deadlier than clade II 1 — but Kindrachuk says that it’s hard to speculate on what that means for humans at this point.
Even when not fatal, mpox can trigger fevers, aches and painful fluid-filled skin lesions.
Growing mpox outbreak prompts WHO to declare global health emergency
Although many reports state that 10% of clade I infections in humans are fatal, infectious-disease researcher Laurens Liesenborghs at the Institute of Tropical Medicine in Antwerp, Belgium, doubts that this figure is accurate. Even the WHO’s latest estimate of a 3.5% fatality rate for people with mpox in the DRC might be high.
There are many reasons that fatality estimates might be unreliable, Liesenborghs says. For one, surveillance data capture only the most severe cases; many people who are less ill might not seek care at hospitals or through physicians, so their infections go unreported.
Another factor that can confound fatality rates is a secondary health condition. For example, people living with HIV — who can represent a large proportion of the population in many African countries — die from mpox at twice the rate of the general population 2 , especially if their HIV is untreated. And the relatively high death rate among children under age 5 could be partly because of malnutrition, which is common among kids in rural parts of the DRC, Liesenborghs says.
The clade Ib virus has garnered particular attention because epidemiological data suggest that it transmits more readily between people than previous strains did, including through sexual activity, whereas clade Ia mostly comes from animals. An analysis posted ahead of peer review on the preprint server medRxiv 3 shows that clade Ib’s genome contains genetic mutations that seem to have been induced by the human immune system, suggesting that it has been in humans for some time. Clade Ia genomes have fewer of these mutations.
But Liesenborghs says that the mutations and clades might not be the most important factor in understanding how monkeypox virus spreads. Although distinguishing Ia from Ib is useful in tracking the disease, he says, the severity and transmissibility of the disease could be affected more by the region where the virus is circulating and the people there. Clade Ia, for instance, seems to be more common in sparsely populated rural regions where it is less likely to spread far. Clade Ib is cropping up in densely populated areas and spreading more readily.
Jean Nachega, an infectious-disease physician at the University of Pittsburgh in Pennsylvania, says that scientists don’t understand many aspects of mpox transmission — they haven’t even determined which animal serves as a reservoir for the virus in the wild, although rodents are able to carry it. “We have to be very humble,” Nachega says.
Just as was the case during the COVID-19 pandemic, health experts are looking to vaccines to help curb this mpox outbreak. Although there are no vaccines designed specifically against the monkeypox virus, there are two vaccines proven to ward off a related poxvirus — the one that causes smallpox. Jynneos, made by biotechnology company Bavarian Nordic in Hellerup, Denmark, contains a type of poxvirus that can’t replicate but can trigger an immune response. LC16m8, made by pharmaceutical company KM Biologics in Kumamoto, Japan, contains a live — but weakened — version of a different poxvirus strain.
Hopes dashed for drug aimed at monkeypox virus spreading in Africa
Still, it’s unclear how effective these smallpox vaccines are against mpox generally. Dimie Ogoina, an infectious-disease specialist at Niger Delta University in Wilberforce Island, Nigeria, points out that vaccines have been tested only against clade II virus in European and US populations, because these shots were distributed by wealthy nations during the 2022 global outbreak . And those recipients were primarily young, healthy men who have sex with men, a population that was particularly susceptible during that outbreak. One study in the United States found that one dose of Jynneos was 80% effective at preventing the disease in at-risk people, whereas two doses were 82% effective 4 ; the WHO recommends getting both jabs.
People in Africa infected with either the clade Ia or Ib virus — especially children and those with compromised immune systems — might respond differently. However, one study in the DRC found that the Jynneos vaccine generally raised antibodies against mpox in about 1,000 health-care workers who received it 5 .
But researchers are trying to fill in some data gaps. A team in the DRC is about to launch a clinical trial of Jynneos in people who have come into close contact with the monkeypox virus — but have not shown symptoms — to see whether it can prevent future infection, or improve outcomes if an infection arises.
Mpox vaccines have been largely unavailable in Africa, but several wealthy countries have pledged to donate doses to the DRC and other affected African nations. The United States has offered 50,000 Jynneos doses from its national stockpile, and the European Union has ordered 175,000, with individual member countries pledging extra doses. Bavarian Nordic has also added another 40,000. Japan has offered 3.5 million doses of LC16m8 — for which only one jab is recommended instead of two.
Monkeypox in Africa: the science the world ignored
None of them have arrived yet, though, says Espoir Bwenge Malembaka, an epidemiologist at the Catholic University of Bukavu in the DRC. Low- and middle-income nations cannot receive vaccines until the WHO has deemed the jabs safe and effective. And the WHO has not given its thumbs up yet. It is evaluating data from vaccine manufacturers, delaying donors’ ability to send the vaccines.
Even when the vaccines arrive, Bwenge Malembaka says, “it’s really a drop in the bucket”. The Africa Centres for Disease Control and Prevention in Addis Ababa, Ethiopia, estimates that 10 million doses are needed to rein in the outbreak.
Bwenge Malembaka says that the uncertainty over vaccine arrival has made it difficult for the government to form a distribution plan. “I don’t know how one can go about this kind of challenge,” he says. Bwenge Malembaka suspects that children are likely to receive doses first, because they are highly vulnerable to clade I, but officials haven’t decided which regions to target. It’s also unclear how the government would prioritize other vulnerable populations such as sex workers, who have been affected by clade Ib. Their profession is criminalized in the DRC, so they might not be able to come forward for treatment.
Researchers lament that public-health organizations didn’t provide vaccines and other resources as soon as the clade I outbreak was identified, especially given lessons learnt from the 2022 global mpox outbreak. “The opportunity was there a couple months ago to cut this transmission chain, but resources weren’t available,” Liesenborghs says. “Now, it will be more challenging to tackle this outbreak, and the population at risk is much broader.”
Nature 633 , 16-17 (2024)
doi: https://doi.org/10.1038/d41586-024-02793-9
Americo, J. L., Earl, P. L. & Moss, B. Proc. Natl Acad. Sci. USA 120 , e2220415120 (2023).
Article PubMed Google Scholar
Yinka-Ogunleye, A. et al. BMJ Glob. Health 8 , e013126 (2023).
Kinganda-Lusamaki, E. et al. Preprint at medRxiv https://doi.org/10.1101/2024.08.13.24311951 (2024).
Yeganeh, N. et al. Vaccine 42 , 125987 (2024).
Priyamvada, L. et al. Vaccine 40 , 7321–7327 (2022).
Download references
Reprints and permissions
Mapping glycoprotein structure reveals Flaviviridae evolutionary history
Article 04 SEP 24
Farmed fur animals harbour viruses with zoonotic spillover potential
Mysterious Oropouche virus is spreading: what you should know
News Q&A 26 AUG 24
Found: a brain-wiring pattern linked to depression
News 04 SEP 24
The hepatitis C virus envelope protein complex is a dimer of heterodimers
How rival weight-loss drugs fare at treating obesity, diabetes and more
News 03 SEP 24
What accelerates brain ageing? This AI ‘brain clock’ points to answers
News 27 AUG 24
Extreme heat is a huge killer — these local approaches can keep people safe
News 22 AUG 24
The Department of Integrative Biology and Pharmacology (https://med.uth.edu/ibp/), McGovern Medical School at The University of Texas Health Scienc...
Houston, Texas (US)
UTHealth Houston
The Yale Stem Cell Center invites applications for faculty positions at the rank of Assistant, Associate, or full Professor. Rank and tenure will b...
New Haven, Connecticut
Yale Stem Cell Center
Join us at MedUni Vienna to explore the pharmacology of circular and stapled peptide therapeutics targetting the κ-opioid receptor in the periphery.
Vienna (AT)
Medical University of Vienna
The Michigan Neuroscience Institute at the University of Michigan invites applications for tenure-track faculty position at the Assistant Professor.
Ann Arbor, Michigan
University of Michigan; Michigan Neuroscience Institute
Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.
Help | Advanced Search
Title: brain-inspired artificial intelligence: a comprehensive review.
Abstract: Current artificial intelligence (AI) models often focus on enhancing performance through meticulous parameter tuning and optimization techniques. However, the fundamental design principles behind these models receive comparatively less attention, which can limit our understanding of their potential and constraints. This comprehensive review explores the diverse design inspirations that have shaped modern AI models, i.e., brain-inspired artificial intelligence (BIAI). We present a classification framework that categorizes BIAI approaches into physical structure-inspired and human behavior-inspired models. We also examine the real-world applications where different BIAI models excel, highlighting their practical benefits and deployment challenges. By delving into these areas, we provide new insights and propose future research directions to drive innovation and address current gaps in the field. This review offers researchers and practitioners a comprehensive overview of the BIAI landscape, helping them harness its potential and expedite advancements in AI development.
Comments: | 35 pages, 4 figures |
Subjects: | Artificial Intelligence (cs.AI) |
Cite as: | [cs.AI] |
(or [cs.AI] for this version) | |
Focus to learn more arXiv-issued DOI via DataCite |
Access paper:.
Code, data and media associated with this article, recommenders and search tools.
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs .
New citation alert added.
This alert has been successfully added and will be sent to:
You will be notified whenever a record that you have chosen has been cited.
To manage your alert preferences, click on the button below.
Please log in to your account
Bibliometrics & citations, view options, index terms.
Applied computing
Life and medical sciences
Health care information systems
Computing methodologies
Artificial intelligence
Computer vision
Computer vision problems
Image segmentation
Interest point and salient region detections
Video segmentation
Machine learning
Machine learning approaches
Neural networks
Otsu’s thresholding technique for mri image brain tumor segmentation.
MRI image segmentation is very challenging area in medical image processing. It is implemented with the low contract of MRI scan. In terms of certain input features or expert information, the major objective of medical image segmentation is to ...
In this research, an automated method is proposed for Brain tumor classification into four different types which is an important step in brain tumor diagnosis. Most of the recent research studies focus on binomial classification of brain MR image into ...
Image processing has plays vital role in today's technological world. It can be applied in numerous application areas such as medical, remote sensing, computer vision etc. Brain tumor is caused due to formation of abnormal tissues within human brain. ...
Published in.
Hindawi Limited
London, United Kingdom
Other metrics, bibliometrics, article metrics.
Login options.
Check if you have access through your login credentials or your institution to get full access on this article.
Share this publication link.
Copying failed.
Affiliations, export citations.
We are preparing your search results for download ...
We will inform you here when the file is ready.
Your file of search results citations is now ready.
Your search export query has expired. Please try again.
Neuroscientists show how the brain implements responses to unexpected events.
Researchers have discovered how two brain areas, neocortex and thalamus, work together to detect discrepancies between what animals expect from their environment and actual events. These prediction errors are implemented by selective boosting of unexpected sensory information. These findings enhance our understanding of predictive processing in the brain and could offer insights into how brain circuits are altered in autism spectrum disorders (ASDs) and schizophrenia spectrum disorders (SSDs).
The research, published today in Nature, outlines how scientists at the Sainsbury Wellcome Centre at UCL studied mice in a virtual reality environment to take us a step closer to understanding both the nature of prediction error signals in the brain as well as the mechanisms by which they arise.
"Our brains constantly predict what to expect in the world around us and the consequences of our actions. When these predictions turn out wrong, this causes strong activation of different brain areas, and such prediction error signals are important for helping us learn from our mistakes and update our predictions. But despite their importance, surprisingly little is known about the neural circuit mechanisms responsible for their implementation in the brain," explained Professor Sonja Hofer, Group Leader at SWC and corresponding author on the paper.
To study how the brain processes expected and unexpected events, the researchers placed mice in a virtual reality environment where they could navigate along a familiar corridor to get to a reward. The virtual environment enabled the team to precisely control visual input and introduce unexpected images on the walls. By using a technique called two-photon calcium imaging, the researchers were able to record the neural activity from many individual neurons in primary visual cortex, the first area in our neocortex to receive visual information from the eyes.
"Previous theories proposed that prediction error signals encode how the actual visual input is different from expectations, but surprisingly we found no experimental evidence for this. Instead, we discovered that the brain boosts the responses of neurons that have the strongest preference for the unexpected visual input. The error signal we observe is a consequence of this selective amplification of visual information. This implies that our brain detects discrepancies between predictions and actual inputs to make unexpected events more salient" explained Dr Shohei Furutachi, Senior Research Fellow in the Hofer and Mrsic-Flogel labs at SWC and first author on the study.
To understand how the brain generates this amplification of the unexpected sensory input in the visual cortex, the team used a technique called optogenetics to inactivate or activate different groups of neurons. They found two groups of neurons that were important for causing the prediction error signal in the visual cortex: vasoactive intestinal polypeptide (VIP)-expressing inhibitory interneurons in V1 and a thalamic brain region called the pulvinar, which integrates information from many neocortical and subcortical areas and is strongly connected to V1. But the researchers found that these two groups of neurons interact in a surprising way.
"Often in neuroscience we focus on studying one brain region or pathway at a time. But coming from a molecular biology background, I was fascinated by how different molecular pathways synergistically interact to enable flexible and contextual regulation. I decided to test the possibility that cooperation could be occurring at the level of neural circuits, between VIP neurons and the pulvinar," explained Dr Furutachi.
And indeed, Dr Furutachi's work revealed that VIP neurons and pulvinar act synergistically together. VIP neurons act like a switch board: when they are off, the pulvinar suppresses activity in the neocortex, but when VIP neurons are on, the pulvinar can strongly and selectively boost sensory responses in the neocortex. The cooperative interaction of these two pathways thus mediates the sensory prediction error signals in visual cortex.
The next steps for the team are to explore how and where in the brain the animals' predictions are compared with the actual sensory input to compute sensory prediction errors and how prediction error signals drive learning. They are also exploring how their findings could help contribute to understanding ASDs and SSDs.
"It has been proposed that ASDs and SSDs both can be explained by an imbalance in the prediction error system. We are now trying to apply our discovery to ASDs and SSDs model animals to study the mechanistic neural circuit underpinnings of these disorders," explained Dr Furutachi.
This research was funded by the Sainsbury Wellcome Centre Core Grant from the Gatsby Charity Foundation and Wellcome (219627/Z/19/Z and 090843/F/09/Z); a Wellcome Investigator Award (219561/Z/19/Z); the Gatsby Charitable Foundation (GAT3212 and GAT3361); the Wellcome Trust (090843/E/09/Z and 217211/Z/19/Z); European Research Council (HigherVision 337797; NeuroV1sion 616509); the SNSF (31003A 169525); Biozentrum core funds (University of Basel).
Story Source:
Materials provided by Sainsbury Wellcome Centre . Note: Content may be edited for style and length.
Journal Reference :
Cite This Page :
Strange & offbeat.
COMMENTS
Research Freedom Google Brain team members set their own research agenda, with the team as a whole maintaining a portfolio of projects across different time horizons and levels of risk. Google Scale As part of Google and Alphabet, the team has resources and access to projects impossible to find elsewhere. ... Papers Accepted to NIPS, 2017
Google Research & Lichtman Lab, Harvard University / D. Berger (rendering) ... The data set itself and a preprint version of this paper were released in 2021. Brain atlases come in many forms ...
Researchers built a 3D image of nearly every neuron and their connections within a small piece of human brain tissue. The left image shows excitatory neurons and the right image shows inhibitory neurons.These versions are shaded according to the size of the neurons' cell bodies (central core), which range from 15-30 micrometers across.
The Google Research team developed advanced AI tools to construct an interactive 3D model of the brain tissue. The model underscores how complex the human brain is: describing just this small sample — one-millionth of the total human brain and about 3 mm long — requires more than a million Gigabytes of data: 1.4 Petabytes.
A major hypothesis in modern neuroscience is that neuron-to-neuron connectivity structure in the brain can be linked to function -- how the brain encodes memories, extracts features from perceptual stimuli, and makes decisions. However, the structure of these brain networks has remained largely unknown, due to technical difficulties involved in ...
Google researchers recently unveiled the largest, most detailed map of the human brain yet.It described just 1 cubic millimeter of brain tissue — the size of half a grain of rice — but at high enough resolution to show individual neurons and their connections to each other, and required 1.4 petabytes of data to encode.. Although it's only a tiny sliver of the brain, the map led to ...
Additionally, in 2016, we welcomed the first cohort of the Google Brain Residency Program, a one-year program for people who want to learn to do machine learning research. In its inaugural year , 27 residents conducted research alongside and under the mentorship of Brain team members , and authored more than 40 papers that were accepted in top ...
In 2021, Google Research and collaborators released the H01 dataset, a 1.4 petabyte rendering of a small sample of human brain tissue. This was the first-ever human connectome project , which used tools and software developed at Google Research to map a small part (roughly one cubic millimeter) of the human brain — all easily accessible with ...
Abstract. The dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention ...
Although the sample represents only one millionth of the volume of a human brain, mapping it in such detail generated 1.4 million gigabytes of data. To manage such a large dataset, the Lichtman Lab teamed with Google's connectomics research group. The collaboration yielded the largest connectomics map of a human sample produced so far.
Collectively, these maps represent more than a decade of human brain mapping research and encompass phenotypes including the first principal component of gene expression 7, 36 neurotransmitter ...
Posted by Jeff Dean, Google Senior Fellow, on behalf of the entire Google Brain Team The Google Brain team works to advance the state of the art in artificial intelligence by research and systems engineering, as one part of the overall Google AI effort. Last year we shared a summary of our work in 2016. Since then, we've continued to make progress on our long-term research agenda of making ...
Posted by Jeff Dean, Google Senior Fellow, on behalf of the entire Google Brain Team The Google Brain team works to advance the state of the art in artificial intelligence by research and systems engineering, as one part of the overall Google AI effort. In Part 1 of this blog post, we shared some of our work in 2017 related to our broader research, from designing new machine learning ...
Google Scholar provides a simple way to broadly search for scholarly literature. Search across a wide variety of disciplines and sources: articles, theses, books, abstracts and court opinions.
Publications. Our teams aspire to make discoveries that impact everyone, and core to our approach is sharing our research and tools to fuel progress in the field. Google publishes hundreds of research papers each year. Publishing our work enables us to collaborate and share ideas with, as well as learn from, the broader scientific community.
Advancing the state of the art. Our teams advance the state of the art through research, systems engineering, and collaboration across Google. We publish hundreds of research papers each year across a wide range of domains, sharing our latest developments in order to collaboratively progress computing and science. Learn more about our philosophy.
Google Brain was a deep learning artificial intelligence research team that served as the sole AI branch of Google before being incorporated under the newer umbrella of Google AI, a research division at Google dedicated to artificial intelligence.Formed in 2011, it combined open-ended machine learning research with information systems and large-scale computing resources. [1]
1. Educational Neuroscience (Teaching for the Brain and Teaching about the Brain) Educational neuroscience is an interdisciplinary field exploring the effects of education on the human brain and promotes the translation of research findings to brain-based pedagogies and policies [].The brain is the target organ of education.
The human brain is the command centre for the nervous system and enables thoughts, memory, movement, and emotions by a complex function that is the highest product of biological evolution. Maintaining a healthy brain during one's life is the uppermost goal in pursuing health and longevity. As the population ages, the burden of neurological ...
Google DeepMind researchers are presenting more than 80 new papers at ICML this year. As many papers were submitted before Google Brain and DeepMind joined forces, papers initially submitted under a Google Brain affiliation will be included in a Google Research blog, while this blog features papers submitted under a DeepMind affiliation.
RNNs have in recent years become the typical network architecture for translation, processing language sequentially in a left-to-right or right-to-left fashion. Reading one word at a time, this forces RNNs to perform multiple steps to make decisions that depend on words far away from each other. Processing the example above, an RNN could only ...
Discover — Discover our latest breakthroughs. See how we're shaping the future. Hear how AI can transform our world. Blog — Discover our latest AI breakthroughs, projects, and updates. Events — Meet our team and learn more about our research. The Podcast — Uncover the extraordinary ways AI is transforming our world on Google DeepMind: The Podcast.
Previous studies have shown that regions of the human brain that are the last to mature, such as parts of the frontal lobe, are the first to show signs of ageing 2, a theory known as 'last in ...
Gateway to research with smart searches, audio podcasts from papers, easy uploads, and interactive chats.
Similar to Pang, Langdon, an associate research scholar in the lab of PNI associate professor Tatiana Engel, Ph.D., will use the Innovator Award funds to support his work on using recurrent neural networks to bridge our understanding of brain circuit connectivity, neural activity, and behavior.
In brain-computer interface (BCI) decoding, attention mechanisms can accurately identify important features in electroencephalogram (EEG) signals, thereby improving decoding accuracy and efficiency.
Article PubMed Google Scholar ... Found: a brain-wiring pattern linked to depression. News 04 SEP 24. ... Research articles
Current artificial intelligence (AI) models often focus on enhancing performance through meticulous parameter tuning and optimization techniques. However, the fundamental design principles behind these models receive comparatively less attention, which can limit our understanding of their potential and constraints. This comprehensive review explores the diverse design inspirations that have ...
The purpose of this paper is to create and put into practice a system for classifying different types of MRI images of brain tumor samples. As a result, this paper concentrated on the tasks of segmentation, feature extraction, classifier building, and classification into four categories using various machine learning algorithms.
The research, published today in ... Group Leader at SWC and corresponding author on the paper. To study how the brain processes expected and unexpected events, the researchers placed mice in a ...