Oculomics - any use for ME/CFS?

[Sasha]

Just watched a clip about 'oculomics' on today's edition of Click, the BBC's technology programme (the episode isn't available on iPlayer yet).

Oculomics is basically looking at large numbers of retinal scans to correlate features of the retina with health and disease. Given that the retina is the only part of the brain that's easily accessible, that's a big deal, and applying machine learning to the scans seems to be yielding results. The Click clip featured research based at Moorfields Eye Hospital in London showing that Parkinson's can be diagnosed 7 years earlier by using this method.

Wondering if it could be of use for ME/CFS and LC? Another use for DecodeME data, since the work featured on Click was done cross-referencing millions of retinal scans at Moorfields with other NHS data on those patients to see if they later developed various diseases, no matter where they moved to in the country. DecodeME has NHS tags, I think? And loads of us will have retinal scans from opticians or hospital eye departments?

@Chris Ponting @Simon M

[Edited to add] The Click episode is now on iPlayer, here, and the oculomics bit starts at 24:34.

 

[Braganca]

Have wondered same.. think it would be a really interesting study. There are lots of new diagnostics in development that could be looked at for ME including this one and breath or sebum tests.

 

[Yann04]

@forestglip this might interest you.

You were talking about plans for a machine learning model to predict PEM based on facial expressions if I remember correctly.

 

[forestglip]

@forestglip this might interest you.

You were talking about plans for a machine learning model to predict PEM based on facial expressions if I remember correctly.

Oh yeah, thanks that's very cool. My idea wasn't working very well, probably because the difference in my face was way too small for it to be picked up by a camera and/or the ML model.

But I'm still always trying to think of a similar concept that might work - training with data before and after large energy changes in myself. If I had access to that kind of equipment, I'd definitely want to see if retinal images could be used for PEM prediction.

 

[Sasha]

The Click episode is now on iPlayer, here, and the oculomics bit starts at 24:34.

 

[Creekside]

A view of my retinas would probably just result in "Boy, this guy has lousy eyes." I recently had a small retinal tear laser welded. Some people got eyeballs; I got lemons.

I do agree that comparing a large number of retinal scans and comparing with controls might reveal something useful, and is relatively inexpensive, since the data already exists.