New Analyzer: Add the Balloon model in differential form

Also due to selfish reasons: since implementing this model as a
Monitor would imply to modify the structure of the Monitors and
probably the Simulator, my evil plan consists in adding the Balloon
model as an analyzer that will operate on Time-Series dtypes and will
make use of the integrators.

My motivation is to have at least a documented version of the Balloon
model. This, even if it is not an ideal form, permits to compare
results obtained via the differential and the convolution-based BOLD
models.

Take the equations from:

Stephan, K. E.; Weiskopf, N.; Drysdale, P. M.; Robinson, P. A. &
Friston, K. J. Comparing hemodynamic models with DCM. (2007)

Results are BOLD Time-Series.