This topic is about applying ML techniques to various aspects of PD.
Regards,
Katherine
https://www.linkedin.com/pub/katherine-moss/3/b49/228
1 Like
All,
I have a B.S. in English and a Masters in C.S. I am a software developer. I have worked several years in biotech. I have worked with databases, with imaging, and with laboratory equipment. I am earning ML specializations in Coursera/DeepLearning.AI right now.
PD has hit my family hard. I would be interested in any aspect of fighting the disease, especially a cure. Most recently I have been reading into the link between bacterial genera that generate sulfur compounds, which then travel to the brain and aggregate. I also found interesting some studies of drawings of waves and spirals created by PD patients vs. healthy patients.
Regards,
Katherine Moss
https://www.linkedin.com/pub/katherine-moss/3/b49/228
1 Like
Hi all,
My name is dinesh and i have a bachelor’s degree in Electronics and communication engineering. I worked as a software developer for 3 years. I have plunged into the world of data science a year ago. Learning though diverse resources and participating in kaggle competitions.I am by no means an expert but just intense passion towards data science to solve some real-world problems.
Looking forward to learn and contribute to the best of my abilities.
Regards
Malloju Dinesh
2 Likes
Hi,
My name is Charles. I am a disabled veteran with Parkinsons. I am 52 and clueless about tech. I am retired from my nurses career taking care of paralyzed vets. Switching careers now cause of my tremor. Taking my first A.I. course so I can still have something to contribute to the world. Would like to be able to find a alternative to the keyboard but then again my voice is shakey. So firsst thing I have to learn is getting used to a new way to interface with a.i.
2 Likes
All,
I looked further into the study of hand drawn images created by both PD patients and healthy patients. The idea is that PD patients draw spirals and waves differently from healthy patients. I have heard that this technical is being used now clinically for diagnosis, but I could not find any other information on it.
There are images available of these drawings on Kaggle. The authors used data augmentation to create some of the samples. The dataset is 264 MB and downloads into two folders (‘healthy’ and ‘parkinson’), with 1,632 drawings in each folder. If you are interested, the dataset is available on Kaggle here: Parkinson’s Disease Augmented Data of Handdrawings (kaggle.com)
The data available on Kaggle does not include demographics of the patients. I was wondering to myself if things like age would make a difference. For example, should a model be trained on various age groups, so that an older person, without PD, but with less hand motor control, be incorrectly marked as having PD.
Katherine Moss
https://www.linkedin.com/pub/katherine-moss/3/b49/228
Hi @Katherine_Moss ,
I am no expert in PD. My thinking is that PD is not necessary age related, there are young PD patients too. Understanding what is going in the brain of PD patients is important. From what I read, the diagnosis of PD is based on a variety of observations/factors. You can trained your model to establish a signature exhibited by PD patient’s motor control, but it is very risky to make a diagnosis based on this one factor.
Kic,
You make good points. I assume that health care professionals would not use just a drawing test and that patients should not either. Thank you.
For those interested, here is a sample of images of spirals created by PD patients. I find the variety of spirals interesting, from a data scientist’s point of view.
Katherine Moss
I have been reading an interesting PD paper. Some bacteria in the gut have been associated with PD, including the Desulfovibrio bacteria. Last May, scientists in Helsinki fed some nematode worms gut material from PD patients. They also fed another group of nematodes gut material from people without PD. Then the scientists did some imaging of the worms’ brains to see if there was a difference. The worms fed material from PD patients showed more Desulfovibrio in the brain than the other worms. The images above show the difference.
From a patient’s point of view, this is an interesting study. From a data science point of view, I am fascinated that one could use images (other than MRI’s) as part of an A.I. project for Parkinson’s Disease. Here is the paper, if you are interested.
Katherine Moss
This is an interesting paper to read as well.
New hope for Parkinson’s disease treatment: Targeting gut microbiota - PMC (nih.gov)
Desulfovibrio bacteria are more abundant in the feces of PD patients compared to healthy people. According to Wikipedia, this species of bacteria is commonly found in aquatic environment with high level of organic material, as well as water logged soils. It is known that PD patients suffer from constipation. So could there be a link here! By detecting/monitoring the level of Desulfovibrio bacteria presence and improving gut health, PD might never develop. Besides, gut health is linked to a whole host of illnesses. If more attention is focused on improving gut health, diet and exercise regimes, we might be able to stay healthy, enjoying healthier aging and better quality of life.
Kic,
You make good points. It is hopeful to think that diet, etc., can make a difference in PD. Thank you for the link. That paper is a bit easier to read than other research papers on the topic.
Thank you,
Katherine Moss
1 Like
All,
Hope you are well.
I found an exciting paper today about a Belgian team’s work. They used fecal matter transplant in a study. It reversed PD symptoms after about 6 months.
Transplanting Feces From One Person to Another Could Ease Parkinson’s Symptoms : ScienceAlert
Katherine
Always keep in mind that correlation does not prove causation.
TMosh,
We must always remember that correlation does not prove causation. Thank you.
However, in the study outlined in the paper I gave a hyperlink for, the lead neurologist, Dr. Arnout Bruggeman from Ghent University said…
“After twelve months, participants who received the healthy donor stool transplant showed a significant improvement in their motor score, the most important measure for Parkinson’s symptoms.”
Some caveats to consider here are that the study group was small, and we don’t know the demographic details of the participants; e.g., the participants might come from the same general genetic pool, and genetics is somehow a factor in success. Here is the original paper, if anyone is interested: Safety and efficacy of faecal microbiota transplantation in patients with mild to moderate Parkinson’s disease (GUT-PARFECT): a double-blind, placebo-controlled, randomised, phase 2 trial - eClinicalMedicine (thelancet.com)
Back to data science for a moment, I wonder how we as data scientists can participate here? What related datasets might be available to us?
Katherine Moss