In plain matlab, you cannot represent such a number as it's way outside the range that a double can represent. The exponent part of a double is 11 bits, offset by 1023, so the minimum number that can be represented by a double is around 2^-1023 (the true minimum is 2^-1074), your number is 2^-4096!
You would have to use vpa from the Symbolic math toolbox (I know nothing about it, I don't have it), or go through Java's BigDecimal, or look through the file exchange for an arbitrary precision type.
