Using AI to fight ALS: Scientist takes on devastating disease to help people – The Augusta Chronicle

After his father died young, Bert Timm of Augusta thought his time might be short, too. So when he got the chance to retire at 55 after 34 years as a postal carrier, he jumped at it and stocked his garage with fishing gear and golf equipment. But six years into enjoying his new life, he got the devastating diagnosis of a progressive neurodegenerative disease. The avid outdoorsman now needs a motorized wheelchair for mobility.
“I worked hard,” said Timm, who never once called in sick. “I feel like I got hoodoo-ed.”
His condition is amyotrophic lateral sclerosis or ALS. It is a relentless disease that attacks the motor neurons of the body, eventually robbing patients of movement, speech, the ability to swallow and eventually the ability to breathe. 
Patients have only a few approved drugs to turn to that can extend their lives by a year or two at most, but there is no cure. Even what causes most cases is unknown.
The urgent need to find new discoveries for patients like Timm is why scientists like Dr. Eric Vitriol at Augusta University are turning to artificial intelligence to help speed up their search for what exactly happens at the cellular level.
Working with mathematicians while he was the University of Florida before coming to Augusta in February, Vitriol and others used machine learning to help create a program that could read highly detailed fluorescent microscope images of a key protein that makes up muscle cells and helps them respond to signals. Machine learning is computer algorithms that can be taught to recognize patterns from data and improve over time based on its own experience.
“When I learned about machine language (programs), I became thrilled because now we could use artificial intelligence to analyze the images and potentially find new things that I wouldn’t have even been able to identify myself,” Vitriol said.
The problem with many of those programs is they use a “black box” approach – they provide answers but not how they arrived at the answers, even when they work perfectly, he said.
His Florida colleague, Dr. Peter Bubinek, specializes in a type of math that seeks to pull out patterns from data. Their machine-learning program not only finds the patterns but quantifies where and why they are different. Another advantage to the program is how quickly it can be trained. Other image-classifying algorithms can take hundreds of images to properly train them but theirs uses far fewer, about 20-30 at a time. That was a practical consideration for Vitriol, whose work can require eight minutes to produce a single large and highly detailed image.
“If I need thousands of those, that’s not going to happen,” he said. “That’s not even feasible.”
More: Augusta clinic, MCG student focus on patient, family in treatment of ALS
When differences between images are obvious, finding those patterns is easier. 
“Other times we actually have to take a lot of images and take measurements in order to see a difference because it is pretty hard even for the human eye to detect,” Vitriol said. And that is where the program can be most helpful when looking at images over time.
“What we would like to do in the future is be able to identify things that are almost invisible in people who are healthy individuals or people who are just getting sick so we can predict what is going to happen in the future,” Vitriol said. “These would be the causative events.”
It is finding those early changes that will make a difference, he said.
“As a cell biologist, I firmly believe that if you can crack the mystery then you can rationally design treatments to address it,” Vitriol said. “And I think the field is moving toward that. We need to find the early causative events because those are the ones we need to prevent instead of only looking at the endpoint and trying to address some of the things that are happening there.” At that point “it’s irreversible,” he said.
As a microscopist, the data he deals with are images and similar to the way other computer algorithms find patterns in data, “being able to see patterns through the noise,” and helping him find small changes to the protein, potentially subtle changes that could have bigger meaning over time, Vitriol said.
It is that way with the disease for many patients. ALS often begins subtly and it may not be apparent anything in the body is changing, as happened with perhaps the most famous person to succumb to it.
“If you think about Lou Gehrig, he had some of the best neuromuscular connections in human history,” Vitriol said of the slugging first baseman for the fearsome New York Yankees of the late 1920s, a lineup often referred to as Murderers’ Row. He died of ALS two years after its effects forced him to retire from baseball.
It began that way for Timm a few years after his retirement when he noticed his left ankle was often stiff. His father’s death at age 57 not long after retiring from the Augusta Fire Department influenced his own retirement decision.
“He lasted less than a year and died of a massive heart attack,” Timm said. “So I figured I will get out at 55 where I can get at least two good years.”
After taking care of his mother who had dementia and died a couple of years later, he pursued his outdoor passions.
“I like to fish and to golf,” Timm said. “I wasn’t much of a housecat, you know.”
The ankle kept bothering and sometimes the left leg would stiffen up, too. His doctor thought it could be the Achilles tendon and prescribed stretching. But it would get worse.
After Hurricane Irma blew through in 2017 and left a bunch of debris on his roof, Timm was up there clearing the gutters when “I realized I couldn’t get up,” he said. “I didn’t have the strength in the left leg.” He managed to crawl over to the garage and haul himself down the ladder. From there, he went from doctor to doctor, test to test, before being referred at AU and Dr. Michael Rivner, an expert in neurodegenerative diseases and ALS.
“I knew that was bad news,” Timm joked. In May 2018, he got the diagnosis of ALS. In just three years, he went from needing a walking cane to a walker to finally, last year, the wheelchair.
He sold the bass boat and the truck to build a series of ramps from the back door to the garage.
“All of that side was fishing stuff,” Timm said, gesturing with his right arm toward the garage outside. “That was my big hobby. I had oodles of stuff. I had rods and reels, all top of the line stuff. So I sold all of that stuff. People come every now and then and shop. I call it Timm’s Outdoors.”
He sits at a kitchen table beneath a wall of mounted trophy fish on plaques, a couple of big crappie, a largemouth bass and a hybrid bass. Most of the time Timm is in the living room where a police scanner constantly squawks. And he can now watch television shows from the 1990s that he’d missed.
“I never watched any of that because I was always working or outside doing stuff,” Timm said. “I’m catching up. They ain’t reruns to me. They’re fresh.”
He hates being indoors but he is slowly losing the use of his outdoor wheelchair, where he could at least go out and pick up the pine cones off his yard or use the leaf blower.
But there isn’t a trace of self-pity in Timm’s voice when he talks about it.
“I get up in the morning, I don’t think, ‘I’ve got ALS,’^” Timm said. “I don’t get thoroughly depressed. I just move on.”
He makes an adventure out of working around the things he can no longer do.
“The best thing I figured out was Door Dash,” he said, of the food delivery service. He has small scissors to cut open ketchup packets he can no longer tear open and he can still microwave things, though it takes an elaborate maneuver around his small kitchen.
There is one aspect of him, a side that takes little coaxing, that shows no sign of diminishing.
“I can still run my mouth with the best of them,” he said, grinning.
Having met ALS patients, Vitriol knows they don’t have time for him to work slowly.
“You realize this is a heart-breaking, devastating illness for people,” he said. “They’re desperate. It does add some urgency, I would say.”

source
Connect with Chris Hood, a digital strategist that can help you with AI.

Leave a Reply

Your email address will not be published. Required fields are marked *

© 2021 AI Caosuo - Proudly powered by theme Octo