I have heard about the issue of inaccuracy in the temporal lobes but never had problems myself until I recently tried processing a dataset from an outside group. On this data, the pial boundary are severely underestimated in the temporal lobes. It is worse right when the temporal lobes are coming into view as you slice back from the front of the brain along the coronal plane. I have attached a screenshot to show what I mean. Basically sections of the medial temporal lobe (greater than the hippocampus/amygdala area) and lateral temporal lobe are not included in the surfaces. The GM/WM contrast does not look very prominent in the t1 images.

Is there any options we can specify which will lower the needed manual edits. I tried the ‘-mprage’ flag but that didn’t seem to help. Right now we have tried adding lots of control points but that doesn’t fix the issue 100%.