• All threads and posts regarding Coronavirus COVID-19 can be found in our area specifically for Coronavirus COVID-19 discussion.

    You can directly access this area >here<.

Is Artifical Intelligence a viable support tool?

KoalaBear

New member
Jan 5, 2020
1
0
I am doing research for an EPQ on how Artificial Intelligence could be the key to supporting Alzheimer's patients, but I know the experience is different for everyone. I have found several start ups, however I would like to know personal opinions from carers, families and the sufferers themselves. If anyone would be willing to give their opinion on the use of technology to support patients verses actual human interaction (or even your own ideas of how AI could help!), it would be greatly appreciated!
 

ally66

New member
Oct 4, 2017
2
0
My father was diagnosed 5 years ago and is now needing 24hour car but we kept him independent longer with the use of cameras and room sensors. He now has a robotic cat whom he loves and interacts with. His speech is not good now but I use the u tube channel on tv and play songs most of the day.
 

DesperateofDevon

Registered User
Jul 7, 2019
2,742
0
Aged mother gets confused with the lifeline box & button. Also of the generation that has no broadband !
confabulations with the radio/ tv can’t imagine what would happen with AI.
 

WJG

Registered User
Sep 13, 2020
51
0
This is more about diagnosis of dementia - I wonder whether AI could be taught how to diagnose the different dementias. For instance, AI could interpret MRI, CT and SPECT scans as well as analysing neuro-psychological and cognitive tests. If patients are being given tests that essentially measure responses, then these can all be done by AI. At its simplest I was surprised that a psychiatrist used paper and pen to score a cognitive test with me recentury - doing this on a laptop or tablet would have been quicker, easier and more accurate.
 

imthedaughter

Registered User
Apr 3, 2019
552
0
This is more about diagnosis of dementia - I wonder whether AI could be taught how to diagnose the different dementias. For instance, AI could interpret MRI, CT and SPECT scans as well as analysing neuro-psychological and cognitive tests. If patients are being given tests that essentially measure responses, then these can all be done by AI. At its simplest I was surprised that a psychiatrist used paper and pen to score a cognitive test with me recentury - doing this on a laptop or tablet would have been quicker, easier and more accurate.
I imagine it could. Some types of dementia are seen on scans by atrophy of certain areas of the brain and AI could be used to 'read' scans and refer for confirmation diagnosis.
The same could be said for the cognitive tests, I would think a GP would have to enter the answers but it could measure some extra things like time taken to respond, number of prompts given, and give a score at the end.
 

WJG

Registered User
Sep 13, 2020
51
0
My understanding is that GPs know little about rare dementias in particular. So why not automate the diagnostic process. Diagnosis of any medical condition is, after all, a matter of algorithms. AI is bound to have a major effect on medicine.
 

Jaded'n'faded

Registered User
Jan 23, 2019
1,107
0
High Peak
I don't think you can automate the process.

Scans don't always show 'evidence' of dementia. Tests are not always conclusive, especially for the rarer dementias where memory is not particularly affected. Use of AI to interpret scans would lead to a lot of people with dementia not being diagnosed.
 

Mydarlingdaughter

Registered User
Oct 25, 2019
134
0
North East England UK
I am doing research for an EPQ on how Artificial Intelligence could be the key to supporting Alzheimer's patients, but I know the experience is different for everyone. I have found several start ups, however I would like to know personal opinions from carers, families and the sufferers themselves. If anyone would be willing to give their opinion on the use of technology to support patients verses actual human interaction (or even your own ideas of how AI could help!), it would be greatly appreciated!

When my Mum was living on her own at home she gradually became unable to use any of the electrical appliances we take for granted, the radio, the television, the CD player, the fridge, and lastly, which was very difficult, the landline telephone. I had of course bought her a basic mobile and shown her how to use it but it was impossible.
I tried to keep in touch by phoning her but after a minute or two for some reason she was unable to hear me. As I could her her I am sassuming that she wasn't holding the receiver to her ear.
She loves to listen to music and I think that were she to have a usable device in her care home which she could essily use that would improve her quality of life.
 

WJG

Registered User
Sep 13, 2020
51
0
I don't think you can automate the process.

Scans don't always show 'evidence' of dementia. Tests are not always conclusive, especially for the rarer dementias where memory is not particularly affected. Use of AI to interpret scans would lead to a lot of people with dementia not being diagnosed.
With all due respect, that is an assumption. It could turn out to be the other way around - that AI would prove to be better at spotting cases than human experts. I think I'm right in saying that AI has been shown to be better at spotting beast cancer in mammograms than trained radiographers. I'm not saying that this is necessarily the case - but I would suggest that it's an interesting area for research.
 

canary

Registered User
Feb 25, 2014
14,317
0
South coast
I think I'm right in saying that AI has been shown to be better at spotting beast cancer in mammograms than trained radiographers.
With all due respect in return, dementia is not like cancer. If you have cancer then, by de facto, there is a growth that shows on an MRI and the trick is spotting it. With dementia the atrophy produced that shows on an MRI is a secondary affect. The primary affect of the disruption in neural cell connections and the build up of abnormal proteins cannot be seen. Because of this the results of the MRI do not equate entirely to the start and/or progression of the disease. Sometimes there is a lot of damage seen on the scan, but hardly any symptoms and at other times there is considerable disruption to thought patterns and brain function, yet very little is seen on a scan. Some rare forms of dementia do not show anything at all on an MRI and have to have specialist forms of scans to reveal the problem. Also required are neuropsychology testing (there are lots of different tests in this category) that can be given, sometimes a lumber puncture and always the history given by people close to the person being tested. So, you see a scan is not the be all and end all of dementia, but simply one test in the midst of an arsenal of testing. Certainly AI can probably eventually be used for scans, but it does not lend itself to this analysis and I doubt it will replace everything else.


BTW, I am intrigued by your interest in AI - all your posts have been about this subject. Do you come from an AI background?
 

WJG

Registered User
Sep 13, 2020
51
0
With all due respect in return, dementia is not like cancer. If you have cancer then, by de facto, there is a growth that shows on an MRI and the trick is spotting it. With dementia the atrophy produced that shows on an MRI is a secondary affect. The primary affect of the disruption in neural cell connections and the build up of abnormal proteins cannot be seen. Because of this the results of the MRI do not equate entirely to the start and/or progression of the disease. Sometimes there is a lot of damage seen on the scan, but hardly any symptoms and at other times there is considerable disruption to thought patterns and brain function, yet very little is seen on a scan. Some rare forms of dementia do not show anything at all on an MRI and have to have specialist forms of scans to reveal the problem. Also required are neuropsychology testing (there are lots of different tests in this category) that can be given, sometimes a lumber puncture and always the history given by people close to the person being tested. So, you see a scan is not the be all and end all of dementia, but simply one test in the midst of an arsenal of testing. Certainly AI can probably eventually be used for scans, but it does not lend itself to this analysis and I doubt it will replace everything else.


BTW, I am intrigued by your interest in AI - all your posts have been about this subject. Do you come from an AI background?
No, not at all - my background is in news, so I try and keep up with developments in technology. And one of my sons is a programmer. I have done various features on robotics. It just seems to me that AI has enormous potential in medicine. Whether people like it or not Doctors are basically computers: running programs that assess information about patients against known factors and arriving at decisions on diagnosis. Temperature + aching limbs + sudden onset + the sniffles = probably the flu. There are plenty of decision trees for relatively simple medical questions - so I see no reason why AI can't be used for more complex questions.
I'm certain that people won't want to think of medicine in such 'inhuman' terms - but I reckon that AI could be better than humans in many respects, in removing fallibility, subjectivity and prejudice. We are basically 'soft machines', whether we want to admit it or not.
 

Jaded'n'faded

Registered User
Jan 23, 2019
1,107
0
High Peak
Also, an algorithm is only as good as the human who wrote it.

Don't get me wrong - I think AI could be enormously helpful in some areas of medicine. But when you're dealing with a person's cognitive state, it's more about interpretation and understanding of the person's behaviours and individual problems. Every person with dementia is different and AI doesn't cope well with things like that.
 

WJG

Registered User
Sep 13, 2020
51
0
With all due respect in return, dementia is not like cancer. If you have cancer then, by de facto, there is a growth that shows on an MRI and the trick is spotting it. With dementia the atrophy produced that shows on an MRI is a secondary affect. The primary affect of the disruption in neural cell connections and the build up of abnormal proteins cannot be seen. Because of this the results of the MRI do not equate entirely to the start and/or progression of the disease. Sometimes there is a lot of damage seen on the scan, but hardly any symptoms and at other times there is considerable disruption to thought patterns and brain function, yet very little is seen on a scan. Some rare forms of dementia do not show anything at all on an MRI and have to have specialist forms of scans to reveal the problem. Also required are neuropsychology testing (there are lots of different tests in this category) that can be given, sometimes a lumber puncture and always the history given by people close to the person being tested. So, you see a scan is not the be all and end all of dementia, but simply one test in the midst of an arsenal of testing. Certainly AI can probably eventually be used for scans, but it does not lend itself to this analysis and I doubt it will replace everything else.


BTW, I am intrigued by your interest in AI - all your posts have been about this subject. Do you come from an AI background?
The other point I wanted to make is that AI could be good precisely because it could take all of these different tests and results into consideration - applying different weighting so where relevant. AI could analyse everything that is known and suggest diagnoses accordingly, in much the same manner as specialists do now. Basically specialists (neuro- psychologists and neurologists, is that right?) are applying sophisticated decision trees informed by their individual knowledge and experience. AI could synthesise ALL this knowledge and experience.

I'm not suggesting that AI could take over from humans, but I'm sure it could go a long way to helping. It seems tragic that people have to wait so long for a diagnosis of FTD. Maybe AI could bring that long wait down.

Thankyou for so clearly laying out the difficulties with diagnosing FTD. I'm personally at the start of this maze, and finding it VERY daunting.
 

WJG

Registered User
Sep 13, 2020
51
0
I am doing research for an EPQ on how Artificial Intelligence could be the key to supporting Alzheimer's patients, but I know the experience is different for everyone. I have found several start ups, however I would like to know personal opinions from carers, families and the sufferers themselves. If anyone would be willing to give their opinion on the use of technology to support patients verses actual human interaction (or even your own ideas of how AI could help!), it would be greatly appreciated!
Further thoughts: AI could surely be if immense benefit in assisting communication for people who find this difficult. At its simplest predictive text does this already - but imagine if AI could use a range of cues to work out what someone wished to communicate.
 

canary

Registered User
Feb 25, 2014
14,317
0
South coast
The trouble with getting AI to diagnose FTD is that it is so complex. It does not fit with the if this.... then that..... pattern of analysis. There are just too many things to take into consideration. Also, doctors do not use the decision tree modal of analysis to make decisions, although the decision tree method can help come to a diagnosis. Apart from very simple stuff, they use the best fit model of analysis, plus a fair bit of intuition. AI cannot yet do this. No computor has the capacity of a human brain (or anywhere near it) yet.
 

WJG

Registered User
Sep 13, 2020
51
0
The trouble with getting AI to diagnose FTD is that it is so complex. It does not fit with the if this.... then that..... pattern of analysis. There are just too many things to take into consideration. Also, doctors do not use the decision tree modal of analysis to make decisions, although the decision tree method can help come to a diagnosis. Apart from very simple stuff, they use the best fit model of analysis, plus a fair bit of intuition. AI cannot yet do this. No computor has the capacity of a human brain (or anywhere near it) yet.
Sorry, I'm not wanting to disagree for the sake of it, but I take issue over 'intuition'. Intuition is surely the application of knowledge plus experience? Some people are much better than others at applying this than others, and can do it in a way that seems that they're not consciously referring back to the sum total of everything they know - but in reality that's what's going on.
As an artist I may claim to intuitively know what colour to use where, but actually that ability cones from years of practice, trial and error, knowledge of other people's work, knowledge of the colour-wheel etc etc..