search

ColeLab / ColeAnticevicNetPartition

Public release of The Cole-Anticevic Brain-wide Network Partition (CAB-NP)
Other
65 stars 20 forks source link

Glasser to Cole Anticevic subcortical parcels #21

Closed amsinha9 closed 3 years ago

amsinha9 commented 3 years ago

How can we map the Glasser parcels to the Cole Anticevic atlas for the subcortical parcels? In CortexSubcortex_ColeAnticevic_NetPartition_wSubcorGSR_parcels_LR_LabelKey.txt, I see that "NETWORKSORTEDORDER" corresponds to the Glasser parcels and "NETWORK" is the mapping to the Cole Anticevic networks, but after row #360 which should start with the subcortical parcels, it is unclear how what network each subcortical parcel corresponds to. It seems as though there are multiple network assignments for the L/R subcortical parcels?

For example, there is "Visual1-01_L-Accumbens"..and so forth for all of the subcortical parcels and then switches to "Somatomotor-01_R-Accumbens", etc.

jielisaji commented 3 years ago

Hi, the original Glasser parcellation did not extend to the subcortex, hence in CortexSubcortex_ColeAnticevic_NetPartition_wSubcorGSR_parcels_LR_LabelKey.txt the subcortical parcels are marked NA under GLASSERLABELNAME.

The network that each subcortical parcel was assigned to can be found in the NETWORK column (e.g. Visual1-01_L-Accumbens is in the Visual1 network).

The name of each subcortical parcel breaks down as follows: Network-SubcorticalParcelNumber_Hemisphere-Structure, so Visual1-01_L-Accumbens is the first parcel in the subcortex to be assigned to the Visual1 network, and is located in the left accumbens. I hope that helps to clear things up?

amsinha9 commented 3 years ago

That makes sense. I'm still unclear on why there are same subcortical parcels have different labels? For example, there is Cingulo-Opercular-01_L-Accumbens. So does L-Accumbens have multiple network assignments? Or is this different from Visual1-01_L-Accumbens?

jielisaji commented 3 years ago

Hi yes -- Accumbens is the name of one of the subcortical structures (similar to Cerebellum, Thalamus etc.), not the name of one of the parcels! There are multiple parcels making up the Accumbens structure, and each parcel has a functional network assignment (e.g. Visual1, Cingulo-Opercular) .

The structure name is provided to give an idea of where the parcel is located in the brain, similar to how the cortical parcel names are appended with -Ctx to indicate they are cortical parcels.

amsinha9 commented 3 years ago

Right, that makes sense. We've been using Glasser's 360 parcels + FreeSurfer's 19 subcortical parcels (i.e. cerebellum, thalamus, etc.), so I was under the impression that the network mapping here treated each subcortical structure as a whole. It was a misunderstanding on my part. I appreciate the clarification.