Connectomics and Brain disease

[dlee138]

In theory, do you guys think all brain diseases could be diagnosed through connectomics or are there brain diseases you think that are inherently not based upon connectivity nor alter connectivity in the brain and therefore can't be diagnosed in such a fashion?

[sgomezr]

Autoimmune neurological diseases, like Myasthenia gravis, perhaps? I think that in that case, antibodies tests are the best/only way of diagnosis. Also, there are many neurological diseases caused by dysfunctional/mutated versions of proteins, such as Neuromyelitis optica, where aquaporins fail to work correctly. In this case, the damage is localized in specific nerves, but in other diseases, all cell types needing the damaged protein fail, so the failure is general... and not based in connectivity. Or maybe I'm wrong, and there could be a way Connectomics could be used to make the diagnosis, but I can't think of one right now. What do you think?

[indigorose1]

Theoretically, if the resolution were good enough, you could see deformities on the synapse level for many "non-connectomic" diseases, but I can't imagine the resolution ever being that good for work in vivo, but maybe for post-mortem diagnosis.

[SandyaS72]

Connectomics may not be the best mode of diagnosing all disorders, but I think because the brain is a network, all diseases change the network in some way that would be helpful to understand for treatment purposes. But depending on the disorder, there may be faster/less expensive/clearer ways to diagnose that are not network-based.

[ajulian3]

Connectomics data could be best implemented when tracking various psychiatric biomarkers where the disease's spatial origins are clearly marked and standardized. Again, our data for this aspect is incredibly limited so it would be best used in tumor detection and analysis as it solely relies on spatial informatics.

[whock]

I think the answer depends in large part on what types of analyses and concepts count under 'connectomics'. In particular, whether or not dynamics comes into play. I do not think the static connectome, and comparisons thereof between patients v controls, can get at all neurological and psychiatric disorders. It isn't simply an alteration in the brain graph but also a pathological/abnormal shift in the dynamics of activity pattern propagation superimposed on a graph. Of course the graph itself is crucial as it constrains what patterns can appear. But we need to explicitly take into account activity patterns - call it flow, traffic, dynamics, time-varying patterns, etc. If connectomics can be defined to include this as part of its domain (and why shouldn't it be able to? Dynamic graphs models exist) then I think there is a much greater chance. But even then, who knows what we'll discover in 15-20 years? Quantum stuff? All about those paradigms.

[akim1]

I don't think all brain diseases are strictly due to purely defects in connection.

There are a lot of diseases that are more "biochemical" in nature. Protein aggregate diseases is one category--Huntington's, Alzheimer's.

Connectomics might be a way to diagnose it, but I think there are much easier ways. For example, you can look for the mutation in Huntingtin gene for Huntington's disease. The Huntingtin gene is also predictive, which connectomics won't be.