Model offers actionable insights for physicians, could augment clinical decisions in resource-limited areas.

A new artificial intelligence model designed by researchers at Harvard Medical School and National Cheng Kung University in Taiwan could bring much-needed clarity to doctors delivering prognoses and deciding on treatments for patients with colorectal cancer, the second deadliest cancer worldwide.

Solely by looking at images of tumor samples — microscopic depictions of cancer cells — the new tool accurately predicts how aggressive a colorectal tumor is, how likely the patient is to survive with and without disease recurrence, and what the optimal therapy might be for them.

Having a tool that answers such questions could help clinicians and patients navigate this wily disease, which often behaves differently even among people with similar disease profiles who receive the same treatment — and could ultimately spare some of the 1 million lives that colorectal cancer claims every year.

A report on the team’s work is published in Nature Communications.

The researchers say that the tool is meant to enhance, not replace, human expertise.

“Our model performs tasks that human pathologists cannot do based on image viewing alone,” said study co-senior author Kun-Hsing Yu, assistant professor of biomedical informatics in the Blavatnik Institute at HMS. Yu led an international team of pathologists, oncologists, biomedical informaticians, and computer scientists.

“What we anticipate is not a replacement of human pathology expertise, but augmentation of what human pathologists can do,” Yu added. “We fully expect that this approach will augment the current clinical practice of cancer management.”

The researchers caution that any individual patient’s prognosis depends on multiple factors and that no model can perfectly predict any given patient’s survival. However, they add, the new model could be useful in guiding clinicians to follow up more closely, consider more aggressive treatments, or recommend clinical trials testing experimental therapies if their patients have worse predicted prognoses based on the tool’s assessment.

The tool could be particularly useful in resource-limited areas both in this country and around the world where advanced pathology and tumor genetic sequencing may not be readily available, the researchers noted.

The new tool goes beyond many current AI tools, which primarily perform tasks that replicate or optimize human expertise. The new tool, by comparison, detects and interprets visual patterns on microscopy images that are indiscernible to the human eye.

The tool, called MOMA (for Multi-omics Multi-cohort Assessment) is freely available to researchers and clinicians.

Extensive training and testing

The model was trained on information obtained from nearly 2,000 patients with colorectal cancer from diverse national patient cohorts that together include more than 450,000 participants — the Health Professionals Follow-up Study, the Nurses’ Health Study, the Cancer Genome Atlas Program, and the NIH’s PLCO (Prostate, Lung, Colorectal and Ovarian) Cancer Screening Trial.

During the training phase, the researchers fed the model information about the patients’ age, sex, cancer stage, and outcomes. They also gave it information about the tumors’ genomic, epigenetic, protein, and metabolic profiles.

Then the researchers showed the model pathology images of tumor samples and asked it to look for visual markers related to tumor types, genetic mutations, epigenetic alterations, disease progression, and patient survival.

The researchers then tested how the model might perform in “the real world” by feeding it a set of images it had not seen before of tumor samples from different patients. They compared its performance with the actual patient outcomes and other available clinical information.

The model accurately predicted the patients’ overall survival following diagnosis, as well as how many of those years would be cancer-free.

The tool also accurately predicted how an individual patient might respond to different therapies, based on whether the patient’s tumor harbored specific genetic mutations that rendered the cancer more or less prone to progression or spread.

In both of those areas the tool outperformed human pathologists as well as current AI models.

The researchers said the model will undergo periodic upgrading as science evolves and new data emerge.

“It is critical that with any AI model, we continuously monitor its behavior and performance because we may see shifts in the distributions of disease burden or new environmental toxins that contribute to cancer development,” Yu said. “It’s important to augment the model with new and more data as they come along so that its performance never lags behind.”

Discerning telltale patterns

The new model takes advantage of recent advances in tumor imaging techniques that offer unprecedented levels of detail, which nonetheless remain indiscernible to human evaluators. Based on these details, the model successfully identified indicators of how aggressive a tumor was and how likely it was to behave in response to a particular treatment.

Based on an image alone, the model also pinpointed characteristics associated with the presence or absence of specific genetic mutations — something that typically requires genomic sequencing of the tumor. Sequencing can be time-consuming and costly, particularly for hospitals where such services are not routinely available.

It is precisely in such situations that the model could provide timely decision support for treatment choice in resource-limited settings or in situations where there is no tumor tissue available for genetic sequencing, the researchers said.

The researchers said that before deploying the model for use in clinics and hospitals, it should be tested in a prospective, randomized trial that assesses the tool’s performance in actual patients over time after initial diagnosis. Such a study would provide the gold-standard demonstration of the model’s capabilities, Yu said, by directly comparing the tool’s real-life performance using images alone with that of human clinicians who use knowledge and test results that the model does not have access to.

Another strength of the model, the researchers said, is its transparent reasoning. If a clinician using the model asks why it made a given prediction, the tool would be able to explain its reasoning and the variables it used.

This feature is important for increasing clinicians’ confidence in the AI models they use, Yu said.

Gauging disease progression, optimal treatment

The model accurately pinpointed image characteristics related to differences in survival.

For example, it identified three image features that portended worse outcomes:

• Greater cell density within a tumor.
• The presence of connective supportive tissue around tumor cells, known as stroma.
• Interactions of tumor cells with smooth muscle cells.

The model also identified patterns within the tumor stroma that indicated which patients were more likely to live longer without cancer recurrence.

The tool also accurately predicted which patients would benefit from a class of cancer treatments known as immune checkpoint inhibitors. While these therapies work in many patients with colon cancer, some experience no measurable benefit and have serious side effects. The model could thus help clinicians tailor treatment and spare patients who wouldn’t benefit, Yu said.

The model also successfully detected epigenetic changes associated with colorectal cancer. These changes — which occur when molecules known as methyl groups attach to DNA and alter how that DNA behaves — are known to silence genes that suppress tumors, causing the cancers to grow rapidly. The model’s ability to identify these changes marks another way it can inform treatment choice and prognosis.

Recent sleep research reveals unexpected connections between the brain and gut.

Sleep is one of the essential human activities — so essential, in fact, that if we don’t get enough sleep for even one night, we may struggle to think, react, and otherwise make it through the day. Yet, despite its importance for function and survival, scientists still don’t fully understand how sleep works.

Enter Dragana Rogulja, a neurobiologist on a quest to unravel the basic biology of sleep.

As a self-described latecomer to science, Rogulja was drawn to questions she considers “broadly interesting and easy to understand on a basic human level.”

One of these questions is what happens when we sleep.

For Rogulja, an associate professor of neurobiology at the Blavatnik Institute at Harvard Medical School, an intriguing aspect of sleep is the loss of consciousness and awareness it brings as the outside world disappears and the inner world takes over.

In a conversation with Harvard Medicine News, Rogulja delved into the details of her sleep research, which uses fruit flies and mice to explore why we need to sleep and how we disconnect from the world during sleep.

Harvard Medicine News: What are you studying in the context of sleep?

Rogulja: There are two main questions that my lab has been pursuing for the past several years. The first is why sleep is necessary for survival. Why is it that if you don’t sleep, you will literally die after not too long? The other question is how your brain disconnects from the environment when you fall asleep. How are stimuli prevented from reaching your brain during sleep?

Elevating the threshold for sensory arousal is essential for sleep, and we want to understand how that barrier is built around the brain. Sleep is one unified state, but it seems to have multiple components that are regulated through separate mechanisms. We want to understand those mechanisms.

HMNews: How has your research changed how you think about sleep?

Rogulja: For a long time, scientists have been guided by the principle that sleep is of the brain, by the brain, and for the brain. As a result, research has largely focused on the brain regarding why sleep is necessary for survival.

However, we now realize that while sleep may be for the brain, it’s not just for the brain. Sleep is a super old behavior that originated in the earliest animals. These animals had no brain and only a very simple nervous system.

Then, as animals became more complex, these brain-related purposes of sleep evolved. However, researchers have looked at the brains of sleep-deprived animals to try to find a reason why they die, and they haven’t found anything. On the other hand, clinical data show that sleep deprivation in humans leads to all kinds of diseases in the body. To us, this really suggested that sleep is about more than just the brain.

Our research tells us that we need to stop thinking about the brain separately from the body when it comes to sleep. I’m still shocked by how neuroscientists think about the brain as having superiority over the body and being at the top of a hierarchy.

To solve the biggest mysteries in neuroscience, we need to take a more integrated approach, which is what my lab is trying to do for sleep. We have found that we really need to think about the whole body to understand sleep. And it makes sense. When you go to sleep, your muscles relax, your circulation changes. Of course, it’s about the whole body.

HMNews: What tools do you use to study sleep?

Rogulja: Historically, a lot of sleep research has been done on humans, but those experiments tend to be limited and descriptive, because you can’t really do experimentation on humans.

However, over the last two and a half decades, scientists have realized that fruit flies sleep; more recently, we figured out that the genes that regulate sleep in flies are conserved in mice.

When I started my lab, we were only using fruit flies as a model system to study sleep, but we have since established a mouse model as well. Fruit flies allow us to test a lot of hypotheses quickly and do large, unbiased genetic screens, and then we can test what we find out in flies in mice, which, as mammals, are more similar to humans.

HMNews: In your 2020 Cell paper, you tackled why sleep is necessary for survival. What’s the answer?

We found that fruit flies who slept less had shorter lifespans: We saw a correlation where the more sleep the flies lost, the faster they died. Interestingly, the mode of sleep deprivation did not matter. What mattered was the amount of sleep lost.

There seemed to be an inflection point where sleep loss was associated with death, which told us that something specific might be happening in the body instead of general wear and tear.

To investigate this further, we stained different organs in sleep-deprived flies with markers of cell damage. We found that in the gut, there was an increase in oxidizing molecules, and the peak of oxidation correlated with the inflection point where the flies started to die.

We confirmed this finding in sleep-deprived mice. But when we gave sleep-deprived flies antioxidants or turned on antioxidant-producing genes in the gut, we found the flies could survive on little or no sleep, suggesting that the gut is a really important target of sleep.

HMNews: Are there any possible applications for humans?

Our findings suggest that if we can prevent oxidation in the gut, we might be able to counteract the effect of losing sleep. This is important because a lot of diseases are tied to gut dysfunction, and many diseases that arise when you don’t sleep enough may actually be a consequence of gut damage.

We’re now starting to think about how to diagnose gut oxidation due to lack of sleep in humans. We want to design “swallowables” — pills or tablets you could swallow that report the oxidative state of your gut by, for example, changing the color of your feces.

We’re also looking for biomarkers: molecules already circulating in the body that indicate lack of sleep and gut oxidation. My lab’s physicians are profiling sleep-deprived mice to look for such biomarkers.

We already have some molecules that are promising markers for oxidation and seem to decrease with antioxidant treatments. Eventually, it may be possible to design supplements that could be taken orally to reverse gut oxidation due to lack of sleep.

HMNews: You just published a new paper in Cell that explores how the brain disconnects from the environment during sleep. Tell us more.

Until now, we knew almost nothing about this. It wasn’t clear if there is a single place in the brain where all sensory information is attenuated during sleep, or if there are multiple such places. For example, are touch and temperature processed the same way during sleep?

Iris Titos, a postdoctoral researcher in my lab, built a system that can deliver mild, medium, or high levels of vibration to fruit flies. Typically, very few flies wake up when you use low-intensity vibrations, and when you use high-intensity vibrations, almost all the flies react.

Then, we did a large-scale screen to identify genes that control how easily flies wake up — so genes that make flies super easy to wake up, and genes that allow flies to sleep through an earthquake.

HMNews: What did the genetic screen show?

The results of the screen were very interesting. We identified a gene that codes for a molecule called CCHa1. When we depleted CCHa1 in the flies, they woke up very easily — so instead of 20 percent waking up at a particular level of vibration, 90 percent woke up.

However, while CCHa1 is present in both the nervous system and the gut, it was only when we depleted it in the gut that flies were roused more easily.

The cells in the gut that produce CCHa1 are called enteroendocrine cells, and they actually share many characteristics with neurons and can even connect and communicate with neurons. These cells face the inside of the gut, and they sort of “taste” the contents of the gut.

We found that the higher concentration of protein in the diet, the more CCHa1 these gut cells produced. This molecule then travels from the gut to the brain, where it signals to a small group of dopaminergic neurons that also receive information about vibrations.

These neurons produce dopamine, which usually promotes arousal, but in this case suppresses arousal. Vibrations weaken the activity of the dopaminergic neurons, which causes the flies to wake up more easily. CCHa1 produced by the gut essentially buffers the dopaminergic neurons against vibrations, allowing the flies to ignore the environment to a greater degree and sleep more deeply.

We also found that the CCHa1 pathway, while critical for gating mechanosensory information, has no influence on how easily the flies wake up when exposed to heat, suggesting that different sensory modalities such as vibration and temperature can be gated independently.

Finally, we showed that a higher protein diet also improved the quality of sleep in mice, making them more resistant to mechanical disturbances. We are now testing whether a similar signaling pathway is involved in mice.

HMNews: What do these findings tell you?

Well, we know from other research that when animals are starving, they suppress sleep in order to forage. By contrast, when they’re satiated, and especially when they’re satiated with proteins, they tend to sleep more.

Now, we’ve shown that animals sleep more deeply and become less responsive when there’s more protein in the diet. This suggests that if animals don’t need to look for food, they can disconnect from the environment and hide somewhere to sleep, which might be safer.

More broadly, our study implies that dietary choices impact sleep quality. Now we can explore this connection in humans to understand how diet could be manipulated to improve sleep.

HMNews: Is there anything about sleep that you think people often misunderstand?

Rogulja: I think people should be aware that how we feel and what’s going on in our bodies don’t have to be the same. In our research, we found that it’s possible to separate the feeling of sleepiness from the need to sleep — some sleep-deprived animals didn’t necessarily feel sleepy, which we could tell because they didn’t sleep extra to catch up on sleep after the deprivation stopped, but these animals still died from the lack of sleep.

This means that even if we can trick ourselves into not feeling sleepy, the lack of sleep still has negative effects on our bodies — for example, if you take a substance that makes you feel awake, the same amount of oxidation is going to happen in your gut.

People may say they’re OK with only a few hours of sleep a night, but they just mean they can make it through the day. Their bodies are still going to register the lack of sleep. We really cannot tell what’s happening in our bodies as a result of sleep deprivation, and we probably need more sleep than we think we do.

Source: HMS

The brain is a wonderful, mysterious thing: three pounds of soft gelatinous tissue through which we interact with the world, generate ideas and construct meaning and representation. Understanding where and how this happens has long been among neuroscience’s fundamental goals.

In recent years researchers have turned to artificial intelligence to make sense of brain activity as measured by fMRI, turning AI models on the data to understand, with increasing specificity, what people are thinking and what those thoughts look like in their brains.

An interdisciplinary team at UC Santa Barbara pushes those boundaries by applying deep learning to fMRI data to create complex reconstructions of what the study subjects saw.

“There are several projects that try to translate fMRI signals into images, mostly because neuroscientists want to understand how brains process visual information,” said Sikun Lin, the lead author of a paper that appeared in a recent NeurIPS conference in November 2022.

According to Lin, UCSB computer science professor Ambuj Singh and cognitive neuroscientist Thomas Sprague, the resulting images generated by this study are both photorealistic and accurately reflect the original “ground truth” images. They noted that previous reconstructions didn’t create images with the same level of fidelity.

Key to their approach is that in addition to images, a layer of information is added through textual descriptions, a move that Lin said was made to add data to train their deep learning model.

Building on a publicly available dataset, they used CLIP (Contrastive Language-Image Pre-training) to encode objective, high-quality text descriptions that pair with the observed images, and then mapped the fMRI data of those observed images on to the CLIP space.

From there they used the output from the mapping models as conditions to train a generative model to reconstruct the image. The resulting reconstructions came remarkably close to the original images viewed by the subjects — closer, in fact, than any previous attempt to reconstruct images from fMRI data. Studies that have followed, including a notable one from Japan, have outlined methods for efficiently manipulating limited data into clear images.

“One of the main gists of this paper is that visual processes are inherently semantic,” Lin said. According to the paper, “the brain is naturally multimodal,” we use multiple modes of information at different levels to gain meaning from a visual scene, such as what is salient, or the relationships between objects in the scene.

“Using a visual representation only might make it more difficult to reconstruct the image,” Lin continued, “but using a semantic representation like CLIP that incorporates text such as the image’s description, is more coherent with how the brain processes information.”

“The science in this is whether the structure of the models can tell you something about how the brain works,” Singh added. “And that’s what we are hoping to try to find.”

In another experiment, for instance, the researchers found that the fMRI brain signals encoded a lot of redundant information — so much that even after masking more than 80% of the fMRI signal, the resulting 10–20% contained enough data to reconstruct an image within the same category as the original image, even though they didn’t feed any image information into the signal reconstruction pipeline (they were working solely from fMRI data).

“This work represents a true paradigm shift in the accuracy and clarity of image reconstruction methods,” Sprague said. “Previous work focused on extremely simplistic stimuli, because our modeling approaches were much simpler. Now, with these new image reconstruction methods in hand, we can advance our cognitive computational neuroscience experiments toward using naturalistic, realistic stimuli without sacrificing our ability to generate clear conclusions.”

At the moment, the reconstruction of brain data into “true” images continues to be labor intensive and out of reach of ordinary use, not to mention the fact that each model is specific to the person whose brain generated the fMRI data. But it doesn’t stop the researchers from musing on the implications of being able to decode what a person is thinking, right down to the layers of meaning that are hyper specific to each mind.

“What I find exciting about this project is whether it might be possible to preserve the cognitive state of a person, and see how these states so uniquely define them,” Singh said. According to Sprague, these methods would allow neuroscientists to conduct further studies measuring how brains change their representations of stimuli — including in representations of robust, complicated scenes — across task changes.

“This is a critical development that will answer fundamental questions about how brains represent information during dynamic cognitive tasks, including those requiring attention, memory and decision-making,” he said.

One of the areas they are now exploring is finding out what and how much is shared between brains so AI models can be constructed without having to start from zero each time.

“The underlying idea is that the human brain across many subjects share some hidden latent commonalities,” said Christos Zangos, a doctoral student researcher in Singh’s lab. “And based on those, currently I’m working on the exact same framework, but I’m trying to train with a different partition of the data set to see to what extent, using small amounts of data, we could build a model for a new subject.”

Source: UC Santa Barbara

The HP Officejet Pro 8600 is a multi-functional printer that boasts a range of useful features for both home and office environments. Its compact design allows for easy placement in limited spaces, making it an ideal choice for those with small offices or home workspaces.

The HP 8600 ink system is designed to handle high-volume printing tasks and features individual ink cartridges for each color, making it easy to only replace the colors that run out. This feature helps to reduce waste and keep costs down.

The printer also comes with a comprehensive software driver that provides users with a range of customization and optimization options. With support for Wi-Fi and Ethernet connectivity, printing from a range of devices is made simple.

For those who frequently print, the HP 8600 printer ink system is a cost-effective solution. The HP Officejet Pro 8600 ink cartridges are readily available in both standard and high-yield options in from Smart Ink. Installing the cartridges is a straightforward process, making it a hassle-free experience.

The HP 8600 Officejet Pro manual is comprehensive and provides clear instructions for setting up, using, and maintaining the printer. It includes information on the HP 8600 ink cartridges, as well as how to replace them. This manual is an excellent resource for those who are new to the HP Officejet Pro 8600.

In the event that replacement HP Officejet Pro 8600 ink is needed, there are various options available. High-quality HP Officejet Pro 8600 replacement ink cartridges can be easily obtained and are designed to meet the same standards of performance and quality as the original cartridges.

In conclusion, the HP Officejet Pro 8600 is a reliable and efficient printer that provides a range of features and benefits for both personal and professional use. Its compact design, cost-effective ink system, and versatile connectivity options make it an excellent choice for those in need of a versatile printing solution.

Cloud edge environments are the points where devices connect to cloud services and are often the most vulnerable to security breaches, as they are exposed to a variety of threats, including unauthorized access, data leakage, and malware attacks. That’s why they need enhanced and reliable protection. G-Core Labs cloud edge environments’ protection is just what you need to provide multi-layered security, including access control, secure authentication, encryption, threat detection, and timely response.

The importance of protecting cloud environments

Securing cloud-edge environments is becoming increasingly important as more organizations use cloud services and apps, and the use of remote and mobile devices continues to grow. The protection ensures the confidentiality, integrity, and availability of cloud resources and services and protects against evolving threats.

When do you need to take care of cloud edge environment protection?

Every organization that uses cloud services or apps requires similar protection. In particular, it is critical for government bodies, financial institutions, and the healthcare industry. Therefore, it is relevant wherever confidentiality, integrity, and availability of resources must be guaranteed.

How to organize such protection and keep the cloud under control?

Securing cloud-edge environments requires a multi-layered approach that involves a combination of security strategies, technologies, and best practices. G-Core Labs offers several protection steps:

1. Implementing access controls to limit access to cloud resources and services to only authorized users and devices. Access control can include measures like multifactor authentication, for example.
2. Using strong encryption that prevents unauthorized access and guarantees the confidentiality and integrity of data in the cloud.
3. Network segmentation for isolating sensitive data and resources from other parts of networks to limit the impact of a security breach and reduce the risk of threat migration.
4. Implementation of an intrusion detection and prevention system (IDPs).

Applying these and other technologies, G-Core Labs establishes a ‌reliable protection of cloud services and applications, using an individual approach and specific requirements in each case. The company’s specialists determine what security measures are necessary to protect your environment and offer the appropriate functions to ensure that your cloud is under your exclusive control.

The second batch of Kirpi armored personnel carriers manufactured by the Turkish company BMC has already arrived in Ukraine.

The news about the arrival of these combat vehicles was announced by the Ukrainian military on social networks.

The initial plan to supply 200 units of MRAP-class Kirpi armored vehicles was first announced in August 2022. In November 2022, Ukrainian soldiers were already operating the first 50 units.

The new arrival is a bit different from the previous one. The main difference is a remotely controlled combat module, instead of a manually-operated turret with a 12.7mm M2 Browning machine gun.

Defense Express noted that this remote-controlled combat module is ANUBIS, manufactured by the Romanian company Pro Optica S.A. This is an interesting development, because the standard version of Kirpi is equipped with a remote-controlled combat module SARP from the Turkish company Aselsan.

It is difficult to tell which combat module is better. But in any case, remote turrets like these offer much more safety to those who operate them.

Turkey usually prefers weapons manufactured by its domestic producers. Choosing a Romanian sub-contractor could have been motivated by the need to reduce short-term costs and increase the number of manufactured Kirpi vehicles.

For Ukraine, it is good news, because the Turkish armored combat vehicles have been delivered in the shortest possible time.

Kirpi and similar machines have already proven their value on the battlefield due to their ability to perform quick maneuvers as well as their exceptionally good level of protection that allows for saving the lives of troops.

A single Kirpi can accommodate 12 soldiers plus one driver. Its empty weight is 18 tons – most of this weight comes from protective armor. The car also has an automatic fire suppression system. Its maximum driving range reaches 1000km (620 miles) when driving at 60 km/h (37 mph).

You can offer your link to a page which is relevant to the topic of this post.