Is it possible that when we all 'see' a color;our mind actually sees different color?

Pythagorean

It's not a simple task at all. Scientists have to navigate through slices of human brain to map the "human connectome" (Sebastian Seung's project). It's difficult to ask slices of brain how they feel, and it's a difficult task to slice up the brain of someone after you've asked them how they feel. For now, Seung will only be working towards getting a general idea of typical connectivity and he has a lot of work left. This is a completely new area of research. The idea is obvious, but the experimental techniques and technology are currently in the innovation stage. fMRI can be used on live, conscious patients; but it detects blood flow in brain regions, not the activity of individual neurons.

Within that scope, we do have quantitatively similar brains as far as we can tell. I mean, lefthanders will sometimes have things backwards, and obviously damaged/diseased brains will have differences that manifest in complaints about altered (hindered) phenomenology (such as the Sack's story, or the famouse phineas gage, or any number of other medical case studies).

Add to that four more difficulties: 1) astrocytes, which play important and active roles in information processing and are even more difficult to track connectivity with. 2) physical system being supervened on is not just neural connections. Genetic expression and metabolis signaling in neurons plays an equally significant role in day-to-day neural activities and the adjustment of synaptic strength. Another method of cell-to-cell communication that's not well understood is 3) tunneling nanotubes (TNT's) which can mediate communication between gap junctions in cells. Finally, 4) the system has degeneracies (Eve Marder's work) so different connectivity schemes can lead to the same functional outcome as long as they are within acceptable parameters. So, some differences may not matter! So it's hard to know when you're chasing your tail or not.

Most of our insight comes from injury studies, like the Sack's one you just read; Ramachandran had a lot of neat anecdotes from being a neurologist, too. Neural processing and connectivity, on the other hand, are hot, new topics. These systems are not easy to model, they are computationally exhaustive and the parameter setting fiddling is infinite. It's not a trivial problem.

Pythagorean

Hrmm.. it appears that as of Sep 20 2011, they (The Human Connectome Project) have gotten a live scanner, but they're still working it out:

http://www.humanconnectomeproject.org/about/scanner/

On that page you can see all the other caveats that come up, too. One thing they're trying to do is reframe a fiber-based brain coordinate system. You can't just use standard 3D to compare brains to each other, variability in local regions add up to global shifts in the position of other local regions, so comparing local regions between two people becomes difficult in standard 3D; this led to the Talairach and MNI systems:

http://www.talairach.org/about.html
http://www.nil.wustl.edu/labs/kevin/.../mnispace.html

The more local the region, the more arbritrary this becomes, though. Does this neuron shifted to the left in this brain equal the neuron shifted to the right in that brain? It may be that you have to map the total developmental story of the neural system to really know how things happened.

But then, you have the problem of degeneracy again. For instance, SIDS cases are associated with serotonin deficiency. SIDS survivors who have a serotonin (5HT) deficiency are hypothesized to have compensated breathing regulation by relying on GABA neurons that were part of the same system. So now a normal (statistically normal) brainstem, and that of a SIDS survivor will have connections that vary their strengths to emphasize different neuron classes regulating the same function. But then if you were to ask about a different, similar function, there may be significant differences. Not only is there degeneracy in functionality, but a single neuron can encode more than one type of information at the same time.