Brain organoid model shows molecular signs of Alzheimer’s before birth

In a model of human fetal brain development, Emory researchers can see perturbations of epigenetic markers in cells derived from people with familial early-onset Alzheimer’s disease, which takes decades to appear. This suggests that in people who inherit mutations linked to early-onset Alzheimer’s, it would be possible to detect molecular changes in their brains before birth. The results were published in the journal Cell Reports. “The beauty of using organoids is that they allow us to Read more

The earliest spot for Alzheimer's blues

How the most common genetic risk factor in AD interacts with the earliest site of neurodegeneration Read more

Make ‘em fight: redirecting neutrophils in CF

Why do people with cystic fibrosis (CF) have such trouble with lung infections? The conventional view is that people with CF are at greater risk for lung infections because thick, sticky mucus builds up in their lungs, allowing bacteria to thrive. CF is caused by a mutation that affects the composition of the mucus. Rabindra Tirouvanziam, an immunologist at Emory, says a better question is: what type of cell is supposed to be fighting the Read more

Department of Biomedical Informatics

Simpler, more portable ECGs: Emory experts hosting computing challenge

An electrocardiogram or ECG is a basic non-invasive diagnostic tool for cardiologists, which conventionally uses 12 electrodes to gather information about electrical signals in the heart and its rhythms. Emory biomedical informatics specialists are hosting an international computing contest aimed at reducing that number as low as possible, so that future portable or wearable ECG devices can be smaller, more convenient and lower in cost.

“We are challenging the research community and industry to design algorithms that classify a large range of cardiac abnormalities using ECGs with varying numbers of channels,” says co-organizer Gari Clifford, PhD, chair of biomedical informatics at Emory University School of Medicine. “The aim is to determine how low we can go — that is, how many channels of data do we need to make an accurate diagnosis?”

The devices could aid in diagnosing common conditions such as atrial fibrillation or supraventricular tachycardia.

“Reduced-lead ECGs are more accessible than standard twelve-lead ECGs in many parts of the world, and the development of effective open-source algorithms for reading reduced-lead ECGs is key for tackling the growing problem of cardiac events internationally,” says co-organizer Matthew Reyna, PhD, assistant professor of biomedical informatics and pharmacology and chemical biology.

The 2021 PhysioNet/Computing in Cardiology Challenge is titled “Will Two Do? Varying Dimensions in Electrocardiography” and calls for designers to build an algorithm that can classify cardiac abnormalities based on 12, 6, 3 and 2-lead ECGs.

So that participants can try out their algorithms, contest organizers are sharing the world’s largest and most diverse set of publicly available ECG data: over 45,000 recordings from China, Europe, Russia and the USA. A similar amount of data has been hidden for the organizers to test the competitors’ algorithms, and a separate evaluation metric will reflects errors of misdiagnosis.

This year’s contest builds upon previous years; in 2017, the challenge was to classify atrial fibrillation based on a single lead, and last year’s was a challenge to diagnose a variety of cardiac problems using standard 12 leads. Contest participants are invited to submit an abstract describing their algorithm, open-source code for their algorithm and a paper on their work.

The contest culminates in the Computing in Cardiology conference, scheduled for September 12-15 in Brno, Czech Republic. More information about the contest is available at PhysioNet.org and requirements for entry and the schedule are detailed at the PhysioNet/Computing in Cardiology Challenge 2021 site. The initial deadline for applying to enter the contest is April 9, 2021.

The contest is part of PhysioNet, an archive of biomedical computing resources supported by the National Institute of Biomedical Imaging and Bioengineering (R01EB030362). It is being co-sponsored by the Gordon and Betty Moore Foundation, Google and MathWorks. Complementary MATLAB licenses and Google Cloud Platform credits are being made available for this year’s challenge. The sponsors are also making it possible to offer several prizes worth several thousand dollars.

Posted on by Quinn Eastman in Heart Leave a comment

Memory screening using eye-tracking on mobile devices

Investigators at Emory Brain Health Center have developed a platform for evaluating visual memory, while someone views photos for a few minutes on an iPad.

Emory researchers, led by Goizuieta Alzheimer’s Disease Research Center director Allan Levey and biomedical informatics chair Gari Clifford, are working with the company Linus Health to develop the VisMET (Visuospatial Memory Eye-Tracking Test) technology further. Results from the most recent version were published in the journal IEEE Transaction on Biomedical Engineering, and the Emory/Linus team continues to refine the technology.

The goal is to screen people for memory issues, identifying those with mild cognitive impairment (MCI) or Alzheimer’s disease. The task — difficult to call it a test — was designed to be more efficient, easier to administer, and more enjoyable than tests currently used.

“We think this could be a sensitive and specific method for detecting visual memory impairment, and it’s convenient enough for use on a wider scale,” Levey says.

The VisMET technology is based on this observation. When someone with MCI or Alzheimer’s views a photo twice, and the photo has been changed the second time (example: an object in the scene has been removed), their eyes spend less time checking the new or missing element in the photo, compared with healthy people. This is because the regions of the brain that drive visual memory formation, such as the entorhinal cortex and hippocampus, are some of the earliest to deteriorate in MCI or Alzheimer’s.

Currently, when someone is evaluated for memory loss, they get a battery of “paper and pencil” tests to assess verbal memory. Researchers say the alternative of viewing photos on a tablet could be less intimidating for those taking the test, as well as easier to administer and score. The only instruction given to study participants was to enjoy the images.

“The current way memory tests are implemented can be stressful,” says software engineer Alvince Pongos, who is co-first author of the IEEE TBME paper, now at MIT’s McGovern Institute for Brain Research. “The difficulty of standard memory tests can lead to test-givers repeating task instructions many times, and to test-takers being confused and frustrated. If we design simpler tasks and make our tools available in the comfort of one’s home, then we remove barriers allowing more people to engage with their health information.”

Read more

Posted on by Quinn Eastman in Neuro Leave a comment

Detecting heart failure via wearable devices

Cardiology researchers have been eagerly taking up consumer electronic devices that include pulse oximeters. Being able to conveniently measure the level of oxygen in someone’s blood is a useful tool, whether one is interested in sleep apnea or COVID-19.

The news that the new Apple Watch includes a pulse oximeter prompted Lab Land to check in with Amit Shah, an Emory cardiologist who has been experimenting with similar devices to discriminate patients with heart failure from those with other conditions.

Shah, together with Shamim Nemati, now at UCSD, and bioinformatics chair Gari Clifford recently published a pilot study on detecting heart failure using the Samsung Simband. The Simband was a prototype device that didn’t make it to the consumer market, but it carried sensors for optical detection of blood volume changes (photoplethysmography), like on the Apple Watch. 

Heart failure causes symptoms such as shortness of breath and leg swelling, but other conditions such as anemia or lung diseases can appear similarly. The idea was to help discriminate people who might need an examination by echocardiogram (cardiac ultrasound).

Read more

Posted on by Quinn Eastman in Uncategorized Leave a comment

Mapping the cancer genome wilderness

A huge cancer genome project has highlighted how DNA that doesn’t code for proteins is still important for keeping our cells on track.

The Pan-Cancer Analysis of Whole Genomes analyzed more than 2,600 tumors from 38 tissues, looking for causative mutations and patterns. Previous work had concentrated on the regions of the genome that code for proteins, but a significant proportion of cancer patients’ tumors don’t carry known “driver” (causative) mutations in protein-coding regions. So this project went out into what used to be called “junk DNA” or the “dark matter” of the genome.

Emory bioinformatics postdoc Matthew Reyna is the first author of one of 23 papers on the PCAWG project, published Feb. 5 in the Nature family of journals. His paper in Nature Communications looks at mutations in non-coding regions of the genome in tumors, analyzing which biological processes are affected.

Some of these were mutations in the promoters of genes encoding well-known cancer suppressors such as p53, but the project also identified new genes containing cancer-driving mutations. A promoter is the stretch of DNA that tells the cell “make RNA copies starting here”.

Reyna contributed to the project while he was at Princeton, working with Benjamin Raphael, and at Emory as well. More recently, he’s been investigating protein-protein interactions with Haian Fu, Andrey Ivanov and others as part of the Cancer Target Discovery and Development (CTD2) project.

Posted on by Quinn Eastman in Cancer Leave a comment

Probing visual memory at leisure

Emory Brain Health researchers have developed a computer program that passively assesses visual memory. An infrared eye tracker monitors eye movements, while the person being tested views a series of photos.

This approach, relatively unstrenuous for those whose memory is being assessed, is an alternative for the diagnosis of mild cognitive impairment or Alzheimer’s disease. It detects degeneration of the regions of the brain that govern visual memory (entorhinal cortex/hippocampus), which are some of the earliest to deteriorate.

The approach was published in Learning and Memory last year, but bioinformatics chair Gari Clifford discussed the project at a recent talk, and we felt it deserved more attention. First author Rafi Haque is a MD/PhD student in the Neuroscience program, with neurology chair/Goizueta ADRC director Allan Levey as senior author.

Eye tracking of people with MCI and Alzheimer’s shows they spend less time checking the new or missing element in the critical region of the photo, compared with healthy controls. Adapted from Haque et al 2019.

The entire test takes around 4 minutes on a standard 24 inch monitor (a follow-up publication on an iPad version is in the pipeline). Photos are presented twice a few minutes apart, and the second time, part of the photo is missing or new – see diagram above. Read more

Posted on by Quinn Eastman in Neuro Leave a comment

Predict the future of critical care in #STATMadness

Emory is participating in STAT Madness, a “March Madness” style bracket competition featuring biomedical research advances instead of basketball teams. Universities or research institutes nominate their champions, research papers that were published the previous year. It’s like “Battle of the Bands.” Whoever gets the loudest — or most numerous — cheers wins.

Please check out all 64 entries, follow the 2019 STAT Madness bracket and vote here:
https://www.statnews.com/feature/stat-madness/bracket/

Emory’s entry for 2019:
It’s like the “precogs” who predict crime in the movie Minority Report, but for sepsis, the deadly response to infection. Shamim Nemati and colleagues have been exploring ways to analyze vital signs in ICU patients and predict sepsis, hours before clinical staff might otherwise notice.

As landmark clinical studies have documented, every hour of delay in giving someone with sepsis antibiotics increases their risk of mortality. So detecting sepsis as early as possible could save thousands of lives. Many hospitals have developed “sniffer” systems that monitor patients for sepsis, but this algorithm tries to spot problems way before they become apparent.

As published in 2018 in Critical Care Medicine, the algorithm can predict sepsis onset—with some false alarms—four, eight, even 12 hours ahead of time. No algorithm is going to be perfect, but it was better than any other previous sepsis predictor. The technology is headed for additional testing and evaluation at several medical centers, as part of a project supported by the federal Biomedical Advanced Research and Development Authority (BARDA).

You can fill out a whole bracket or you can just vote for Emory. The contest will last several rounds. The first round began on Monday, March 4, and lasts until the end of the week. Before 10 am Eastern time Monday morning, there were already more than 5,000 brackets entered!

If Emory advances, then people will be able to continue voting for us starting on Friday. Emory’s first opponent is a regional rival, Vanderbilt University School of Medicine. We are on the upper left side of the bracket.

STAT News is a Boston-based news organization covering biomedical research, pharma and biotech. If you feel like it, please share on social media using the hashtag #statmadness.

Posted on by Quinn Eastman in Uncategorized Leave a comment

CAPTCHA some cancer cells

Humans are good at deciphering complex images, compared to computers. Until recently, internet users often needed to verify that they were human by completing a CAPTCHA security check. A familiar variety asked the user to check all the boxes that contain a car, or a street sign.

If we asked random people off the street to look at pathology slides and “quick, check all the boxes that contain tumor cells,” what would happen? The accuracy, compared to a trained pathologist, wouldn’t be very good.

Not as easy as labeling which boxes contain street signs!

This challenge of expertise – crowdsourcing and pathology are not immediately compatible – is what Lee Cooper and colleagues sought to overcome in a recent paper published in Bioinformatics. So they put together something they called “structured crowdsourcing.”

“We are interested in describing how the immune system behaves in breast cancers, and so we built an artificial intelligence system to look at pathology slides and identify the tissue components,” Cooper says.

His group was particularly interested in the aggressive form of breast cancer: triple negative. They used pathology slide images from the Cancer Genome Atlas, a National Cancer Institute resource. The goal was to mark up the slides and label which sections contained tumor, stroma, white blood cells, dead cells etc.

They used social media to recruit 25 volunteers — medical students and pathologists from around the world (Egypt, Bangladesh, Saudi Arabia, United Arab Emirates, Syria, USA). Participants underwent training and used Slack to communicate and learn about how to classify images. They collaborated using the Digital Slide Archive, a tool developed at Emory. Read more

Posted on by Quinn Eastman in Cancer Leave a comment

Life-saving predictions from the ICU

It’s similar to the “precogs” who predict crime in the movie Minority Report, but for sepsis, the deadly response to infection. That’s how Tim Buchman, director of the Emory Critical Care Center, described an emerging effort to detect and ward off sepsis in ICU patients hours before it starts to make their vital signs go haywire.

As landmark clinical studies have documented, every hour of delay in giving someone with sepsis antibiotics increases their risk of mortality. So detecting sepsis as early as possible could save lives. Many hospitals have developed “sniffer” systems that monitor patients for sepsis risk. See our 2016 feature in Emory Medicine for more details.

What Shamim Nemati and his colleagues, including bioinformatics chair Gari Clifford, have been exploring is more sophisticated. A vastly simplified way to summarize it is: if someone has a disorderly heart rate and blood pressure, those changes can be an early indicator of sepsis.* It requires continuous monitoring – not just once an hour. But in the ICU, this can be done. The algorithm uses 65 indicators, such as respiration, temperature, and oxygen levels — not only heart rate and blood pressure. See below.

Example patient graph. Green = SOFA score. Purple = Artificial Intelligence Sepsis Expert (AISE) score. Red = official definition of sepsis. Blue = antibiotics. Black + red = cultures.    Around 4 pm on December 20, roughly 8 hr prior to any change in the SOFA score, the AISE score starts to increase. The top contributing factors were slight changes in heart rate, respiration, and temperature, given that the patient had surgery in the past 12hr with a contaminated wound and was on a mechanical ventilator. Close to midnight on December 21, other factors show abnormal changes. Five hours later, the patient met the Sepsis-3 definition of sepsis.

As recently published in the journal Critical Care Medicine, Nemati’s algorithm can predict sepsis onset – with some false alarms – 4, 8 even 12 hours ahead of time. No predictor is going to be perfect, Nemati says. The paper lays out specificity, sensitivity and accuracy under various timelines. They get to an AUROC (area under receiving operating characteristic) performance of 0.83 to 0.85, which this explainer web site rates as good (B), and is better than any other previous sepsis predictor. Read more

Posted on by Quinn Eastman in Heart, Immunology Leave a comment

Big data with heart, for psychiatric disorders

Imagine someone undergoing treatment by a psychiatrist. How do we know the treatment is really working or should be modified?

To assess whether the patient’s condition is objectively improving, the doctor could ask him or her to take home a heart rate monitor and wear it continuously for 24 hours. An app connected to the monitor could then track how much the patient’s heart rate varies over time and how much the patient moves.

Heart rate variability can be used to monitor psychiatric disorders

MD/PhD student Erik Reinertsen is the first author on two papers in Physiological Measurement advancing this approach, working under the supervision of Gari Clifford, interim chair of Emory’s Department of Biomedical Informatics.

Clifford’s team has been evaluating heart rate variability and activity as a tool for monitoring both PTSD (post-traumatic stress disorder) and schizophrenia. Clifford says his team’s research is expanding to look at treatment-resistant depression and other mental health issues.

For clinical applications, Clifford emphasizes that his plans focus on tracking disease severity for patients who are already diagnosed, rather than screening for new diagnoses. His team is involved in much larger studies in which heart rate data is being combined with physical activity data from smart watches, body patches, and clinical questionnaires, as well as other behavioral and exposure data collected through smartphone usage patterns.

Intuitively, heart rate variability makes sense for monitoring PTSD, because one of the core symptoms is hyperarousal, along with flashbacks and avoidance or numbness. However, it turns out that the time that provides the most information is when heart rate is lowest and study participants are most likely asleep, or at their lowest ebb during the night.

Home sleep tests generate a ton of information, which can be mined. This approach also fits into a trend for wearable medical technology, recently highlighted in STAT by Max Blau (subscription needed).

The research on PTSD monitoring grows out of work by cardiologists Amit Shah and Viola Vaccarino on heart rate variability in PTSD-discordant twin veterans (2013 Biological Psychiatry paper). Shah and Vaccarino had found that low frequency heart rate variability is much less (49 percent less) in the twin with PTSD. Genetics influences heart rate variability quite a bit, so studying twins allows those factors to be accounted for. Read more

Posted on by Quinn Eastman in Heart, Neuro Leave a comment