New studies cast doubt on fMRI for pain DX

[drusso]

http://gizmodo.com/new-study-casts-...ium=sharefromsite&utm_source=Gizmodo_facebook

Modern-day phrenology. At least the PhD's had a good ride raking in grant money...

---

Members don't see this ad.

 

[101N]

The "study" was a letter to JAMA http://bit.ly/1N0GzTL

 

[lobelsteve]

The "study" was a letter to JAMA http://bit.ly/1N0GzTL

http://archneur.jamanetwork.com/article.aspx?articleid=2516854

 

[drusso]

The "study" was a letter to JAMA http://bit.ly/1N0GzTL

"The biggest issue with these studies—and many neuroimaging studies—are the limitations of reverse inference, where researchers use types of brain activity, sometimes referred to as “blobs” for the shapes they form on fMRI scans, to deduce that a specific cognitive event like processing pain is occurring. While this is technically not impossible to do, especially with new techniques that apply machine learning, “blobology” is a very tricky business."

Get psuedo-science out of pain medicine. No one should be promoting specious data to clinicians. There is simply not enough certainty to know how higher cognitive functions and emotions relate to fMRI images. It's the intellectually honest thing to do.

 

[Aether2000]

I am working on building a tricorder but am having trouble finding the plans...

 

[lobelsteve]

I am working on building a tricorder but am having trouble finding the plans...

I have most of the parts in my garage.

 

[bedrock]

I am working on building a tricorder but am having trouble finding the plans...

plans were buried with leonard nemoy

 

[Ducttape]

The study doesn't disprove any of the research about pain, into. What it seems to specifically state is that the areas that are thought to be "the pain areas" are not specific or inclusive to pain.

Reading it another way, it might reinforce the concept that pain is not pain alone, but pain includes all the other subjective feelings and emotions that are triggered by these specific regions....


Sent from my iPhone using SDN mobile

 

[pastafan]

http://blogs.scientificamerican.com...-prize-in-neuroscience-the-dead-salmon-study/

 

[drusso]

The study doesn't disprove any of the research about pain, into. What it seems to specifically state is that the areas that are thought to be "the pain areas" are not specific or inclusive to pain.

Reading it another way, it might reinforce the concept that pain is not pain alone, but pain includes all the other subjective feelings and emotions that are triggered by these specific regions....


Sent from my iPhone using SDN mobile

The problem is fMRI research is a lot like the problems with meta-analysis: The data collection can't be experimentally controlled or manipulated. You just can't control a subject's mental events. So, the studies depend upon the "law of large numbers" to "average out" the signal/noise ratio using statistical methods. There is a lot of editorial discretion in applying the statistics to data or judging what even counts as data! At best, fMRI or meta-analysis can allow you do INFER new questions or future experiments, but neither is not a robust enough method to DEDUCE causal relationships. Let alone plan policy.

Still, there is a lot of money on the line for academics to promote the narrative that "fMRI can diagnose what kind of pain problem a person has." So, when any speaker promotes fMRI research as a diagnostic tool (not simply a basic science instrument) they are promoting the diagnosis of medical conditions on the basis of a non-FDA approved diagnostic test. There is no validation process of the method, no proficiency testing, and no cross validation between other labs. The vast majority of these papers are coming of isolated labs. It is absolutely the same as the quacks who stand up in front of patient and provider groups promoting a proprietary test for Lyme Disease--everyone else has it wrong and THEY have the TRUE markers for what constitutes a positive serology for Lyme Disease.

It doesn't take much to defy this logic: All pain is in the brain. fMRI show that brain areas activated by pain and emotional suffering are indistinguishable. Therefore, pain is emotional suffering. You can't substitute any complex process for pain: Anxiety, depression, fear, anger, envy, disgust. What's intellectually dishonest is that I think that people promoting this narrative know that the assumptions are a house of cards.

 

[101N]

Pain tracts are somatotopically mapped in the cord and pain is somatopically mapped in the brain as well. The nociceptive and emotional centers are both involved,
and the extent of one vs another is individual. But, multiple fMRI studies have already confirmed the success of validated screening tools to identify a disproportionate
emotional contribution - catastrophizing, higher VAS, sensory pain descriptors, etc - before the MRI is even done.

The resistance to fMRI findings here is similar to - and from the same people - the resistance to opioid dosing and risk of ODD in the "Prescription Opioid Epidemic In a Nutshell"
thread. "Centralized" or "emotional pain" - just like opioid dosage guidelines - threatens market share for IPM.

 

[nvrsumr]

I had been turned into a "believer" in fibro based on some functional MRI studies when I was a resident. May have to change back to my previous skepticism.

 

[lobelsteve]

I had been turned into a "believer" in fibro based on some functional MRI studies when I was a resident. May have to change back to my previous skepticism.

I believe in firm handshakes. Same pattern as FMS on Fmri.

 

[pastafan]

Cluster failure: Why fMRI inferences for spatial extent have inflated false-positive rates

http://www.pnas.org/content/113/28/7900.full

 

[drusso]

Cluster failure: Why fMRI inferences for spatial extent have inflated false-positive rates

http://www.pnas.org/content/113/28/7900.full

"This calls into question the validity of countless published fMRI studies based on parametric clusterwise inference. It is important to stress that we have focused on inferences corrected for multiple comparisons in each group analysis, yet some 40% of a sample of 241 recent fMRI papers did not report correcting for multiple comparisons (26), meaning that many group results in the fMRI literature suffer even worse false-positive rates than found here (37)."

21st Century phrenology: I can't imagine anyone using fMRI claims to support an argument in light of these findings.

 

[SeniorWrangler]

I spent many months working on a paper about fMRI findings in chronic pain, and this is pretty much what I concluded. fMRI, at best, is like the Cold War science of kremlinology, where they tried to figure out who was really in charge of the nukes based on who stood where during the May Day parade and whether the lights were left on in the Moscow Institute of Nukeology.

 

[clubdeac]

Interesting..... Daniel Amen MD, psychiatrist has made hundreds of thousands and maybe millions off this fMRI premise